DETAILED ACTION
This communication is in response to the Application No. 18/362,200 filed on June 22, 2026
in which Claims 1 - 21 are presented for examination.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
The amendments filed on June 22, 2026 have been considered. Claims 1, 10, 12-14, 16-17
have been amended. No claims have been cancelled. Claim 21 has been added. Thus, Claims 1-21 are pending and presented for examination.
Applicant’s arguments filed June 22, 2026 with respect to the 35 U.S.C. 103 rejections have been considered but are moot because the new ground of rejection.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-21 are rejected under 35 U.S.C. 103 as being unpatentable over Wang et al. (hereinafter Wang) (CN 111882052), in view of Arash et al. (hereinafter Arash) (TW 202029067), in view of Roberto (hereinafter Roberto, a non-patent literature reference titled “A scalable, low-latency, high-throughput, optical interconnect architecture based on arrayed waveguide grating routers”), in view of Xu (hereinafter Xu, a non-patent literature reference titled “11 tera-flop per second photonic convolutional accelerator for deep learning optical neural networks”, and in further view of Brea (hereinafter Brea) (US 7296045).
Regarding Claim 1, Wang teaches an optical computing apparatus (Wang, Claim 1, “A photonic convolutional neural network system, wherein it comprises: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths;”, thus the disclosed photonic convolutional neural network system comprises interconnected optical components that generate, modulate, and process optical signals to perform convolution operations, and therefore constitutes an optical computing apparatus) comprising:
a light source array configured to send a plurality of first groups of first optical signals based on first data to be computed, wherein each of the first groups represents a plurality of elements in the first data (Wang, Page 2 – line 21, “The embodiment of the present invention provides a photonic convolutional neural network system, comprising: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths; a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 3 – line 1, “The photonic convolutional neural network system according to one embodiment of the present invention, the output end of the wavelength division multiplexer comprises a plurality of output channels, each of the output channel corresponding to output one of the wavelength division optical signal; the second modulator has a plurality of sub-modulators, the plurality of sub-modulators are corresponding to the plurality of output channels”, thus a light source array configured to send a plurality of first groups of first optical signals based on first data to be computed, wherein each of the first groups represents a plurality of elements in the first data is disclosed because Wang teaches a multi-wavelength light source that outputs optical signals having a plurality of wavelengths, corresponding to a plurality of optical signals. Wang further teaches that a first modulator loads elements of a first matrix onto the optical signals, such that the optical signals carry the first data, and that the signals are output via multiple channels. Accordingly, the multi-wavelength light source corresponds to the light source array, the plurality of wavelengths corresponds to the plurality of first optical signals, and the loaded elements correspond to the plurality of elements represented by each group)
[…] coupled to the light source array and configured to:
receive the first groups (Wang, Page 2- line 21, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the first groups is disclosed because Wang teaches that the first modulator is in optical communication connection with the multi-wavelength light source and receives the optical signals output therefrom. Since the optical signals include multiple wavelengths carrying elements of the first matrix / the first data, the first modulator necessarily receives the groups of optical signals representing the elements)
modulate, based on second data, the first groups to provide modulated second data (Wang, Page 2- line 21, “The embodiment of the present invention provides a photonic convolutional neural network system, comprising: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths; a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 2 – line 29, “a second modulator, the input end of the second modulator is in optical communication connection with the output end of the wavelength division multiplexer, the second modulator is used for loading the element in the second matrix to be processed to the plurality of wavelength division optical signals to obtain a plurality of second modulation optical signals”, thus modulate, based on second data, the first groups to provide modulated second data is disclosed because Wang teaches that the first modulator loads elements of a first matrix onto optical signals to obtain a first modulated light signal, and that a second modulator further loads elements of a second matrix, corresponding to the second data, onto the optical signals to obtain second modulated optical signals. Since the optical signals carrying the first data are further modulated with additional data, this corresponds to modulating the first groups to produce modulated second data)
output a plurality of intermediate optical signals based on the modulated second data (Wang, Page 5 – line 14, “For example, the second modulator is used for respectively loading a plurality of elements in the second matrix to be processed to a plurality of sub-wave optical signals to obtain a plurality of second modulated light signals”, & Page 5 – line 25, “As shown in FIG. 4, the y1 is loaded to C1, then obtaining x1y1, x2y1 ... x9y1, corresponding y2 is loaded to C2, then obtaining x1y2, x2y2 ... x9y2, and so on y9 is loaded to C9, then obtaining x1y9, x2y9 ... x9y9”, thus output a plurality of intermediate optical signals based on the modulated second data is disclosed because Wang teaches that the second modulator loads elements of the second matrix onto the optical signals to generate a plurality of second modulated light signals. Wang further discloses that this modulation produces multiple outputs (e.g., x1y1, x2y1 … x9y9), which are optical signals resulting from the applied second data. The plurality of second modulated light signals corresponds to a plurality of intermediate optical signals generated based on the modulated second data)
And a wavelength router (Wang, Page 8 – line 12, “Regarding the on-chip planar array waveguide grating, AWG (Arrayed Waveguide Grating) is the preferred technology in dense wavelength division multiplexing system (DWDM). A group of grating formed by array waveguide with equal length difference, using the ability of dividing wave”, Page 8 – line 23, “Therefore, light of different wavelengths in the output flat waveguide is diffracted and focused to different output channel waveguide position; after outputting the output channel waveguide, finishing the wavelength distribution, namely de-multiplexing function. The inverse process of this process, that is, if the signal light is reversely input, the multiplexing function is finished, the principle is the same”, thus a wavelength router is disclosed because Wang teaches an arrayed waveguide grating (AWG) that separates optical signals by wavelength and directs them to different output channel waveguides. Wang also discloses that light of different wavelengths is diffracted and focused to different output channel positions, indicating routing of wavelengths to distinct paths) comprising […] wherein the wavelength router is configured to:
receive the intermediate optical signals […] (Wang, Page 5 – line 36, “the input end of the wavelength division multiplexer 60 is in optical communication connection with the output end of the delay line 50, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus receive the intermediate optical signals is disclosed because Wang teaches that the wavelength division multiplexer is in optical communication connection with the output of the delay line and receives the plurality of second modulated light signals. Since these second modulated light signals correspond to intermediate optical signals, the wavelength division multiplexer receives the intermediate optical signals)
route, based on […] a wavelength of each intermediate optical signal […], the intermediate optical signals as a plurality of second optical signals […], wherein intermediate optical signals corresponding to products of elements of the first data and the second data […] (Wang, Page 5, “the input end of the second modulator 40 is connected with the output end of the wavelength division multiplexer 30 optical communication connection, the second modulator 40 for the element to be processed in the second matrix is loaded to a plurality of wavelength division optical signal to obtain a plurality of second modulated light signal”, & Page 7, “the y1, y2 ... y9 are loaded to C1, then obtaining the x1y1, x1y2 ... x1y9, the y1, y2 ... y9 are loaded to C2, then obtaining the x2y1, x2y2 ... x2y9, and so on, y2 ... y9 are loaded to C9, then obtaining x9y1, x9y2 ... x9y9”, & Page 8, “for different wavelengths of light, the phase difference is different, Therefore, light of different wavelengths in the output flat waveguide is diffracted and focused to different output channel waveguide position; after outputting the output channel waveguide, finishing the wavelength distribution, namely de-multiplexing function”, thus route, based on […] a wavelength of each intermediate optical signal […], the intermediate optical signals as a plurality of second optical signals […], wherein intermediate optical signals corresponding to products of elements of the first data and the second data […] is disclosed because Wang teaches that wavelength-division optical signals
C
1
through
C
9
carry elements
x
1
through
x
9
of the first matrix, corresponding to the first data. Wang further teaches loading elements
y
1
through
y
9
of the second matrix, corresponding to the second data, onto the wavelength-division optical signals to produce second-modulated optical signals having values
x
i
y
j
. The second-modulated optical signals therefore correspond to the intermediate optical signals representing products of elements of the first data and the second data. Wang also teaches that light of different wavelengths is diffracted and focused to different output-channel waveguide positions, such that the wavelengths of the intermediate optical signals are used to route the intermediate optical signals as a plurality of second optical signals)
And output the second optical signals to indicate a computing result of the first data and the second data (Wang, Page 9 – claim 1, “a photoelectric detector, the input end of the photoelectric detector is in optical communication connection with the output end of the wavelength division multiplexer, for sampling the combined wave light signal and converting into electric signal, outputting the convolution result”, & Page 6 – line 2, “It can be understood that, as shown in FIG. 5, the photoelectric detector 70 can select the proper sampling time, sampling the information carried on the combined wave optical signal, for example, according to the sampling time of the obtained first matrix X and the second matrix Y by bit multiplying result, namely only collecting the x1y1 in the wave-combining optical signal. x2y2 ... x9y9 information, and converting it into electric signal, finally obtaining the convolution result”, thus output the second optical signals to indicate a computing result of the first data and the second data is disclosed because Wang teaches that the photoelectric detector samples the combined optical signal carrying results of operations between elements of the first matrix and the second matrix (e.g., x1y1, x2y2 … x9y9) and outputs a convolution result. Since these optical signals encode the results of the computation between the first and second data, they correspond to second optical signals indicating a computing result)
Wang does not explicitly teach a modulator array, a plurality of input ports coupled to the modulator array and a plurality of output ports, receiving at the plurality of input ports, and […] to the plurality of output ports[…], […] based on both a wavelength […] and an identity of one of the plurality of input ports at which […] is received […]are routed to the same output port, […], […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data, and […] with an equal sequence number difference (i-j) […].
However, Arash teaches:
A modulator array (Arash, Page 36 – line 1, “The modulator array 144 is configured to receive light input from the laser unit 142, and modulate the intensity of the received light input based on the modulator control signal (which is an electrical signal). Examples of modulators include Mach-Zehnder interference (MZI) modulators, ring resonator modulators, and electro-absorption modulators”, thus a modulator array is disclosed because Arash teaches a modulator array configured to receive light input and modulate the intensity of the received light based on control signals)
a plurality of input ports coupled to the modulator array and a plurality of output ports (Arash, Page 10 – line 25, “an input waveguide array for receiving a light input vector; an optical interference unit, which is in optical communication with the input waveguide array, for performing a conversion of the light input vector into a second optical signal array Linear conversion; and the output waveguide array, in optical communication with the optical interference unit, for guiding the second optical signal array, wherein at least one input waveguide in the input waveguide array passes through the optical interference unit and each output waveguide in the output waveguide array Optical communications”, thus a plurality of input ports coupled to the modulator array and a plurality of output ports is disclosed because Arash teaches an input waveguide array for receiving light input and an output waveguide array for guiding optical signals. Since waveguides function as optical input and output interfaces, the input waveguide array corresponds to a plurality of input ports and the output waveguide array corresponds to a plurality of output ports)
[…] to the plurality of output ports[…] (Arash, Page 40-41, “The OMM unit 150 may include an array of input waveguides 152 to receive light input vectors; an optical interference unit 154 in optical communication with the array of input waveguides 152; and an array of output waveguides 156 in optical communication with the optical interference unit 154. The optical interference unit 154 linearly converts the light input vector into a second optical signal array. The array of output waveguides 156 guides the second array of optical signals output by the optical interference unit 154. At least one input waveguide in the array of input waveguides 152 optically communicates with each output waveguide in the array of output waveguides 156 through the optical interference unit 154. For example, for an optical input vector of length N, the OMM unit 150 may include N input waveguides 152 and N output waveguides 156’, thus […] to the plurality of output ports […] is disclosed because Arash teaches that the optical interference unit converts the optical input vector into a second array of optical signals and that an array of N output waveguides guides those signals. The N output waveguides correspond to the plurality of output ports, and guiding the second array of optical signals through the output waveguides corresponds to directing the second optical signals to the plurality of output ports)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Wang’s approach of sending a plurality of optical signals using a light source array and routing the optical signals based on wavelengths using a wavelength router to perform computation with Arash’s approach of using a modulator array coupled to a plurality of input ports and a plurality of output ports for receiving at the plurality of input ports and routing to the plurality of output ports, thereby reducing power consumption by leveraging the modulator array architecture optimized for efficient optical signal modulation (Arash, Page 37 – line 12, “In some embodiments, the modulator of the modulator array 144 and/or the OMM unit 150 can be designed to reduce power consumption, so that when the modulator is operated to generate a modulation value representing a more frequently occurring coefficient It consumes less power while operating, and consumes more power when operating the modulator to generate modulation values representing coefficients that occur less frequently. For example, for certain data sets that are known to have certain characteristics, power consumption can be reduced. Figure 42 shows the modulation value probability distribution diagram 4200 (dotted line) superimposed on the modulator power diagram 4202 (solid line) for the specific design of the modulator and/or OMM unit 150 of the modulator array 144”, & Page 38 – line 4, “The power distribution shown in Figure 42 shows that the zero modulation power is used to achieve the zero modulation value, but in other embodiments, there may be residual low but non-zero modulation power at the zero modulation value. For these low coefficient weighted data sets, the power consumption can usually be reduced by using a modulator”, thereby reducing power consumption by leveraging the modulator array optimized for efficient optical signal modulation is disclosed because Arash teaches that the modulator array can be designed such that modulation values corresponding to more frequently occurring coefficients consume less power, resulting in an overall reduction in power consumption for optical signal processing)
Wang combined with Arash does not explicitly teach […] based on both a wavelength […] and an identity of one of the plurality of input ports at which […] is received […]are routed to the same output port, […], […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data, and […] with an equal sequence number difference (i-j) […].
However, Roberto teaches:
[…]based on both a wavelength […] and an identity of one of the plurality of input ports at which […] is received […]are routed to the same output port, […] (Roberto, Page 1 – Section 1, “An arrayed waveguide grating router (AWGR) [7], [8] is an example of devices with such wavelength routing capability. As Fig. 1(a) and (b) illustrate, the well-known wavelength routing property of an AWGR allows any input port to communicate with any output port simultaneously using different wavelengths without contention”, & Page 2 – Fig 1 a and b,
PNG
media_image1.png
412
489
media_image1.png
Greyscale
,
& Page 8 – Section V, “Second, we show a scenario with network contention where C1 and C2 send packets to C3 with random arrival time. To avoid resynchronization in the FPGA GTH receiver, tunable optical delay lines in each transmission path guarantee clock-phase matching of the packets at AWGR output 3”, thus […] based on both […] a wavelength […] and an identity of one of the plurality of input ports at which […] is received […] are routed to the same output port […] is disclosed because Roberto teaches an AWGR in which the destination output port is determined by the combination of the wavelength used and the input port through which the signal enters, as illustrated by the wavelength-routing property and wavelength-assignment table of Figures 1(a) and 1(b). Roberto further teaches that packets originating from C1 and C2 are directed to the same AWGR output 3)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Wang and Arash with Roberto by incorporating Roberto’s AWGR wavelength-routing technique into the optical computing system of Wang and Arash. Wang teaches generating wavelength division optical signals that represent products of elements of first and second matrices, while Arash teaches communicating optical signals through arrays of input and output waveguides. Roberto teaches that an AWGR routes signals between input and output ports based on wavelength and permits simultaneous optical communication without contention. Therefore, a POSITA would have been motivated to use Roberto’s AWGR to route Wang’s wavelength encoded product signals from Arash’s input waveguides to selected output waveguides based on the wavelength of each signal and the input port through which the signal is received. This would allow multiple product signals to be routed in parallel, reduce contention, and improve the speed, scalability, and energy efficiency of the optical convolution system (Roberto, Page 1 - Section I, “Compared to electrical interconnects, optical interconnects provide (1) higher transmission bandwidth and lower energy consumption independently of distance, (2) inherent parallelism, (3) low interference and crosstalk, and (4) low parasitic” and “the well-known wavelength routing property of an AWGR allows any input port to communicate with any output port simultaneously using different wavelengths without contention”)
Wang and Arash combined with Roberto does not explicitly teach […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data, and […] with an equal sequence number difference (i-j) […].
However, Xu teaches:
[…] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data (Xu, Page 3, “The input data vector X is encoded as the intensity of temporal symbols in a serial electrical waveform at a symbol rate 1/τ (baud), where τ is the symbol period. The convolution kernel is similarly represented by a weight vector W of length R that is then encoded in the optical power of the microcomb lines through spectral shaping performed by a Waveshaper. The temporal waveform X is then multi cast onto the kernel wavelength channels via electro-optical modulation, thus generating the replicas weighted by W. Next the optical waveform is transmitted through a dispersive delay with a delay step (between adjacent wavelength channels) equal to the symbol duration of X, effectively achieving time and wavelength interleaving. Finally, the delayed and weighted replicas are summed via high speed photodetection so that each time slot yields a convolution between X and W for a given convolution window, or receptive field. As such, the convolution window effectively slides at the modulation speed matching the baud rate of X. Each output symbol is the result of R multiply-and-accumulate operations”, thus […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data is disclosed because Xu teaches that the input vector X, corresponding to the first data, is encoded using optical intensity and that the convolution kernel W, corresponding to the second data, is encoded using the optical power of respective wavelength channels. Xu further teaches generating weighted optical replicas representing products of elements of X and W, aligning and summing those replicas through photodetection, and producing an output symbol for each time slot. Because each output symbol results from the multiply-and-accumulate operations for one convolution window, each output symbol represents one element of the convolution result. In the combined system, applying this intensity-based convolution technique to the plurality of output ports results in the light intensity at each output port representing a respective element of the convolution result)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to further combine Wang, Arash, and Roberto with Xu by incorporating Xu’s intensity based optical convolution and photodetection technique into the combined optical routing system. Wang teaches producing optical signals corresponding to products of elements of two matrices, Arash teaches guiding optical signals through a plurality of output waveguides, and Roberto teaches routing the optical signals to selected output ports. Xu teaches encoding input data and convolution weights using optical intensity and optical power, generating weighted optical replicas, and summing those replicas through high-speed photodetection so that each output symbol represents the multiply-and-accumulate result for a convolution window. Therefore, a POSITA would have been motivated to apply Xu’s photodetection technique to the product signals routed to each output port so that the signals received at that port are summed and represented as a respective convolution-output element. This would provide a direct and high speed technique for obtaining convolution results from the routed optical product signals and would take advantage of the parallel wavelength channels already used by the combined system (Xu, Page 3, “Finally, the delayed and weighted replicas are summed via high speed photodetection so that each time slot yields a convolution between X and W for a given convolution window, or receptive field. As such, the convolution window effectively slides at the modulation speed matching the baud rate of X. Each output symbol is the result of R multiply-and-accumulate operations , with the computing speed given by 2R/τ FLOPS. Since the speed of this process scales with both the baud rate and number of wavelengths, it can be dramatically boosted into the Tera-FLOP regime by using the massively parallel wavelength channels of the microcomb source”)
Wang, Arash, and Roberto combined with Xu does not explicitly teach […] with an equal sequence number difference (i-j) […].
However, Brea teaches:
[…] with an equal sequence number difference (i-j) […] (Brea, Par. [0192], “Each output vector sample y.sub.p is produced by the time reversal of the matrix sequence {C.sub.p,p=0,1,2, . . . ,M-1} and its circular shifting to the right. The resulting operation shall be called circular matrix convolution. Note that the sum of the indices of each sample product inside the summation modulo M is equal to the index of the output being calculated by the circular matrix convolution”, thus Brea teaches grouping product terms according to a common convolution-output index. Because one of the sequences is time-reversed before the products are formed, the sum-of-indices relationship for the reversed sequence corresponds to a difference
i
j
between the original sequence numbers. Product terms having the same sequence-number difference therefore contribute to the same convolution-output element)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to further combine Wang, Arash, Roberto, and Xu with Brea by using Brea’s known convolution indexing relationship to determine which optical product signals should be routed to the same output port. Wang teaches optical signals representing products of first data and second-data elements, Roberto teaches routing optical signals according to wavelength and input-port identity, and Xu teaches summing product signals to produce individual convolution-output elements. Brea teaches that the product terms contributing to a convolution output are grouped according to a common output index and that one sequence is time reversed before the product terms are formed. Therefore, a POSITA would have been motivated to assign the wavelengths and input ports of the product signals according to Brea’s convolution index relationship so that products having the same sequence number difference (i-j) are directed to the same output port. Xu’s photodetection technique could then sum the grouped product signals at that port to produce the corresponding convolution element. This would implement the known mathematical organization of convolution products directly in the optical routing hardware and allow multiple convolution elements to be calculated efficiently and in parallel (Brea, Paragraph [0192], “Each output vector sample y.sub.p is produced by the time reversal of the matrix sequence {C.sub.p, p=0,1,2, . . . ,M-1} and its circular shifting to the right. The resulting operation shall be called circular matrix convolution. Note that the sum of the indices of each sample product inside the summation modulo M is equal to the index of the output being calculated by the circular matrix convolution”)
Regarding Claim 2, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Wang further teaches:
[…] one of the second optical signals comprising at least two intermediate optical signals, and wherein each of the at least two intermediate optical signals indicates a product of an element in the first data and an element in the second data (Wang, Page 6-7 – line 44, “It can be understood that the plurality of elements in the first matrix is not loaded on each wavelength of the optical signal according to the time sequence, but one element is loaded on each wavelength, for example: λ1 x1 λ9 loading x9, at this time 9 of the wavelength is loaded with 9 elements. the second modulator is used for loading multiple elements in the second matrix to be processed to each wavelength division optical signal according to the time sequence so as to obtain multiple second modulation optical signals”, & Page 7 – line 4, “for example, the y1, y2 ... y9 are loaded to C1, then obtaining the x1y1, x1y2 ... x1y9, the y1, y2 ... y9 are loaded to C2, then obtaining the x2y1, x2y2 ... x2y9, and so on, y2 ... y9 are loaded to C9, then obtaining x9y1, x9y2 ... x9y9. Similarly, the photoelectric detector only collecting the x1y1, x2y2 ... x9y9 information in the wave-combining optical signal, and converting it into electric signal, finally obtaining the convolution result”, thus one of the second optical signals comprising at least two intermediate optical signals, and wherein each of the at least two intermediate optical signals indicates a product of an element in the first data and an element in the second data is disclosed because Wang teaches that elements of the first matrix are loaded onto different wavelengths and that elements of the second matrix are subsequently loaded onto those wavelength-specific optical signals to generate products (e.g., x1y1, x1y2 … x9y9). Wang also discloses that each wavelength carries multiple such product terms, indicating that a given optical signal comprises multiple intermediate optical signals, each representing a product of an element from the first data and an element from the second data)
Wang does not explicitly teach […] wherein a first output port in the output ports is configured to output […].
However, Arash teaches […] wherein a first output port in the output ports is configured to output […] (Arash, Page 116 – line 9, “The computing device of claim 17, wherein the replication module includes at least one replication module having an optical splitter, and the optical splitter transmits a predetermined ratio of a power of a light wave at an input port of the replication module To a first output port of the copy module, and send a remaining proportion of the power of the light wave to a second output port of the copy module at the input port of the copy module.”, thus wherein a first output port in the output ports is configured to output is disclosed because Arash teaches an optical splitter within a replication module that receives a light wave at an input port and distributes the light wave to multiple output ports, including a first output port, thereby causing the first output port to output a portion of the light wave)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Wang’s approach of one of the second optical signals comprising at least two intermediate optical signals, wherein each of the intermediate optical signals indicates a product of an element in the first data and an element in the second data, with Arash’s approach of configuring a first output port in a plurality of output ports to output an optical signal, in order to split and distribute the second optical signals comprising the intermediate optical signals to a plurality of output ports (Arash, Page 68- line 20, “the 1×2 optical amplitude modulator includes a ring resonator 2222 configured to divide the optical power of the optical signal from the input port 2221 to two output ports”). Arash further teaches that the modulator array can be designed to reduce power consumption during optical signal modulation (Arash, Page 37 – line 12, “In some embodiments, the modulator of the modulator array 144 and/or the OMM unit 150 can be designed to reduce power consumption, so that when the modulator is operated to generate a modulation value representing a more frequently occurring coefficient It consumes less power while operating, and consumes more power when operating the modulator to generate modulation values representing coefficients that occur less frequently. For example, for certain data sets that are known to have certain characteristics, power consumption can be reduced. Figure 42 shows the modulation value probability distribution diagram 4200 (dotted line) superimposed on the modulator power diagram 4202 (solid line) for the specific design of the modulator and/or OMM unit 150 of the modulator array 144”, & Page 38 – line 4, “The power distribution shown in Figure 42 shows that the zero modulation power is used to achieve the zero modulation value, but in other embodiments, there may be residual low but non-zero modulation power at the zero modulation value. For these low coefficient weighted data sets, the power consumption can usually be reduced by using a modulator”, thereby reducing power consumption by leveraging the modulator array optimized for efficient optical signal modulation, and thereby providing an additional benefit of improved energy efficiency)
Regarding Claim 3, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Wang further teaches:
wherein each of the elements corresponds to one of the first optical signals of one wavelength, and wherein each of the first groups comprises first optical signals of different wavelengths (Wang, Page 4 – line 39, “the number of wavelengths of the optical signal output by the multi-wavelength light source 10 according to the number of elements in the first matrix, the number of wavelengths in this example can be 9, then the wavelength corresponding to the elements in the first matrix is respectively represented as λ1 λ9, The first modulator 20 simultaneously loads each element of the first matrix X to 9 wavelengths in the form of light intensity, and the signal loading interval of x1 to x9 is Δ”, & Page 5 – line 1, “In other words, each wavelength is loaded with 9 elements, on each wavelength, the signal loading interval of x1 to x9 is Δ t1, namely, the first modulated light signal can be composed of 9 wavelengths, each wavelength is loaded with 9 elements according to the time sequence”, thus wherein each of the elements corresponds to one of the first optical signals of one wavelength, and wherein each of the first groups comprises first optical signals of different wavelengths is disclosed because Wang teaches that the number of wavelengths corresponds to the number of elements in the first matrix, such that each element is associated with a respective wavelength. Wang further teaches that the optical signal includes multiple wavelengths, and that each wavelength carries signals corresponding to elements, indicating that the first groups include optical signals of different wavelengths)
Regarding Claim 4, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Wang further teaches:
wherein the wavelength router is an arrayed waveguide grating router (AWGR) or an etched diffraction grating router (EDGR) (Wang, Page 3 – line 23, “A photonic convolutional neural network system according to one embodiment of the present invention, the wavelength division multiplexer comprises: sheet upper plane array waveguide grating, cascade Mach-Zehnder interferometer or cascade micro-ring resonant cavity”, & Page 8 – line 12, “Regarding the on-chip planar array waveguide grating, AWG (Arrayed Waveguide Grating) is the preferred technology in dense wavelength division multiplexing system (DWDM). A group of grating formed by array waveguide with equal length difference, using the ability of dividing wave. The principle is: after the multiplexing signal light containing a plurality of wavelengths is output by the central input channel waveguide, diffraction occurs in the input flat waveguide, the input concave grating is performed with power distribution, and coupled into the array waveguide area”, thus wherein the wavelength router is an arrayed waveguide grating router (AWGR) or an etched diffraction grating router (EDGR) is disclosed because Wang teaches that the wavelength division multiplexer may be implemented using an arrayed waveguide grating, and further explains that an AWG is used to separate optical signals based on wavelength through diffraction and routing to different waveguides. Since an AWG performs wavelength-based routing of optical signals, it corresponds to a wavelength router, including an arrayed waveguide grating router)
Regarding Claim 5, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Arash further teaches:
a detector array comprising a plurality of detectors and configured to detect light intensities of the second optical signals to obtain the computing result (Arash, Page 79 – line 36, “In this embodiment, the optical amplitude modulator array can be replaced by a detector array to convert optical signals into electrical signals, followed by electronic subsystems (such as ASIC, processor, or SoC). Optionally, if optical signal routing is to be used for a summation module configured to detect optical signals, the electronic subsystem may include the use of an array of electrically-modulated optical sources ) Of electro-optical conversion”, & Page 46 – line 3, “The input data set processed by the ANN computing system usually includes data with a resolution greater than 1 bit. For example, ordinary pixels of a gray-scale digital image may have an 8-bit resolution, that is, 256 different levels. One way to represent and process the data in the optical domain is to encode 256 pixels with different intensity levels as 256 different power levels of the optical signal input to the OMM unit 150. Optical signals are analog signals in nature, so they are easily affected by noise and detection errors”, thus a detector array comprising a plurality of detectors and configured to detect light intensities of the second optical signals to obtain the computing result is disclosed because Arash teaches a detector array configured to convert optical signals into electrical signals, indicating a plurality of detectors arranged in an array for detecting optical signals. Arash further teaches that optical signals are encoded using different power levels corresponding to different intensity levels, and that such signals are subject to detection, thereby indicating that the detector array detects light intensities of the optical signals to obtain the computing result. Incorporating Arash’s detector array is consistent with Arash’s modulator array optimized for efficient optical signal modulation, which reduces overall power consumption)
Regarding Claim 6, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Arash further teaches:
a light emitting array configured to separately emit the first optical signals of different wavelengths based on the elements, wherein each of the first optical signals is used to indicates one element in the first data (Arash, Page 9-10 – line 45, “The laser unit can be configured to generate multiple wavelengths. The optical modulator may include: a complex optical modulator group configured to generate a complex optical input vector, each optical modulator group corresponds to a wavelength and generates a corresponding optical input vector with a corresponding wavelength; and an optical multiplex The device is configured to combine the light input vector into a combined light input vector including the wavelength”, & Page 77 – line 14, “FIG. 35C shows an embodiment of the system configuration 3500 for the implementation of the photoelectric matrix multiplication unit 3520 of wavelength division multiplexing for performing vector matrix multiplication using a 2×2 element matrix, in which a summation operation is performed in the electrical domain. In this embodiment, the input vector is [Image Omitted] And the matrix is [Image Omitted] . In this embodiment, the input vector has multiple wavelengths λ1, λ2, and λ3, and each element of the input vector is encoded on a different optical signal. Two different copy modules 1902 perform optical copy operations to separate calculations on different paths”, thus a light emitting array configured to separately emit the first optical signals of different wavelengths based on the elements, wherein each of the first optical signals is used to indicate one element in the first data is disclosed because Arash teaches a laser unit configured to generate multiple wavelengths and a plurality of optical modulator groups, each corresponding to a respective wavelength and generating a corresponding optical signal. Arash further teaches that each element of the input vector is encoded on a different optical signal having a different wavelength, thereby indicating that optical signals of different wavelengths are generated for respective elements and correspond to one element each)
And an optical splitting device (Arash, Page 21 – line 21, “In some embodiments, the optical splitter may include a waveguide optical splitter that transmits a predetermined proportion of the power of the light wave guided by the input optical waveguide to the first output optical waveguide, and transfers the light wave guided by the input optical waveguide The remaining proportion of the power is sent to the second output optical waveguide”, & Page 64 – line 51, “Optical devices (such as splitters and optical amplitude modulators) and electronic devices (such as optical detectors and operational amplifiers (op-amp)) can be manufactured on a common substrate. Alternatively, different devices with different substrates can be used to implement different parts of the system, and those devices can communicate through communication channels”, thus an optical splitting device is disclosed because Arash teaches an optical splitter that receives a light wave from an input optical waveguide and transmits a predetermined proportion of the optical power to a first output optical waveguide and a remaining proportion to a second output optical waveguide, thereby splitting the optical signal into multiple outputs) coupled to the light emitting array and configured to:
Arash does not explicitly teach receiving the first optical signals and splitting the first optical signals into the first groups.
However, Wang further teaches:
receive the first optical signals (Wang, Page 2 – line 23, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the first optical signals is disclosed because Wang teaches that the first modulator is in optical communication connection with the output of the multi-wavelength light source, such that the optical signals generated by the light source are provided to and received at the input of the first modulator)
split the first optical signals into the first groups (Wang, Page 2 – line 26, “a wavelength division multiplexer, the input end of the wavelength division multiplexer is in optical communication connection with the output end of the first modulator, the wavelength division multiplexer is used for decomposing the first modulation light signal into a plurality of wavelength division optical signals”, & Page 5 – line 7, “The wavelength division multiplexer 30 divides the first modulated light signal composed of a plurality of wavelengths to obtain a plurality of split-wave optical signals, for example, a plurality of split-wave optical signals may be represented by C1 to C9”, thus split the first optical signals into the first groups is disclosed because Wang teaches that the wavelength division multiplexer decomposes the first modulated light signal into a plurality of wavelength division optical signals and divides the signal into multiple split-wave optical signals (e.g., C1 to C9), thereby separating the optical signals into multiple groups)
Regarding Claim 7, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 6 as cited above and Arash further teaches:
wherein the light emitting array comprises a plurality of lasers configured to send the first optical signals based on the elements (Arash, Page 81 – line 19, “However, the laser unit 142 may also include multiple lasers, which can use different optical signals modulated to different corresponding light waves (for example, each with a line width of 1 nm or less) to perform wavelength division multiplexing. Use operation. Different light waves can have peak wavelengths, and the wavelength distance separated from each other is greater than the line width of the individual laser (for example, greater than 1 nm)”, thus wherein the light emitting array comprises a plurality of lasers configured to send the first optical signals based on the elements is disclosed because Arash teaches that the laser unit may include multiple lasers that generate different light waves with different wavelengths, indicating a plurality of lasers emitting optical signals. Arash further teaches that these optical signals are used in wavelength division multiplexing and are modulated to correspond to data, thereby indicating that the plurality of lasers send optical signals based on the elements)
Regarding Claim 8, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 6 as cited above and Arash further teaches:
a plurality of lasers configured to send a plurality of third optical signals of different wavelengths (Arash, Page 81 – line 7, “The fourth modulated optical signal 4134 is copied by the copy module 4144 to generate copies of the optical signals sent to the matrix multiplication modules 4116d and 4118d. The outputs of the matrix multiplying units 4116c and 4116d are combined using the optical coupler 4120b, and the combined signal is detected by the optical detector 4122b. The outputs of the matrix multiplying units 4118a and 4118b are combined using an optical coupler, and the combined signal is detected by an optical detector”, & Page 81 – line 19, “However, the laser unit 142 may also include multiple lasers, which can use different optical signals modulated to different corresponding light waves (for example, each with a line width of 1 nm or less) to perform wavelength division multiplexing. Use operation. Different light waves can have peak wavelengths, and the wavelength distance separated from each other is greater than the line width of the individual laser (for example, greater than 1 nm), thus a plurality of lasers configured to send a plurality of third optical signals of different wavelengths is disclosed because Arash teaches that the laser unit may include multiple lasers that generate optical signals modulated to different corresponding light waves having different wavelengths, thereby indicating that a plurality of lasers emit a plurality of optical signals of different wavelengths)
Arash does not explicitly teach a plurality of modulators coupled to the lasers and configured to:
receive the third optical signals and respectively modulate the elements to the third optical signals to obtain the first optical signals.
However, Wang further teaches:
a plurality of modulators (Wang, Page 3 – line 4, “the second modulator has a plurality of sub-modulators, the plurality of sub-modulators are corresponding to the plurality of output channels; the number of the sub-modulator and the wavelength number of the optical signal, the element number in the first matrix, the element number in the second matrix are equal”, thus a plurality of modulators is disclosed because Wang teaches a second modulator comprising a plurality of sub-modulators corresponding to multiple output channels) coupled to the lasers and configured to:
receive the third optical signals (Wang, Page 2 – line 23, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the third optical signals is disclosed because Wang teaches that the modulator is in optical communication connection with the output of the light source, such that the optical signals generated by the light source are provided to and received at the input of the modulator)
And respectively modulate the elements to the third optical signals to obtain the first optical signals (Wang, Page 2 – line 23, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 4 – line 37, “the number of wavelengths of the optical signal output by the multi-wavelength light source 10 according to the number of elements in the first matrix, the number of wavelengths in this example can be 9, then the wavelength corresponding to the elements in the first matrix is respectively represented as λ1 λ9, The first modulator 20 simultaneously loads each element of the first matrix X to 9 wavelengths in the form of light intensity, and the signal loading interval of x1 to x9 is Δ”, thus respectively modulate the elements to the third optical signals to obtain the first optical signals is disclosed because Wang teaches that the first modulator loads elements of the first matrix onto optical signals output from the light source to obtain first modulated light signals. Wang further teaches that each element is loaded onto a corresponding wavelength of the optical signals in the form of light intensity, thereby indicating that the elements are respectively modulated onto the optical signals to produce modulated optical signals)
Regarding Claim 9, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 6 as cited above and Wang further teaches:
an optical frequency comb source configured to send a third optical signal comprising a plurality of different wavelengths (Wang, Page 3 – line 20, “The photonic convolutional neural network system according to one embodiment of the present invention, the multi-wavelength light source comprises: optical frequency comb, micro-ring resonant cavity or mode-locked laser”, & Page 4 – line 22, “The multi-wavelength light source 10 is used for outputting an optical signal having a plurality of wavelengths”, thus an optical frequency comb source configured to send a third optical signal comprising a plurality of different wavelengths is disclosed because Wang teaches that the multi-wavelength light source may comprise an optical frequency comb and is configured to output an optical signal having a plurality of wavelengths, thereby indicating that the optical frequency comb source sends an optical signal including multiple different wavelengths)
a wavelength division demultiplexer (Wang, Page 2 – line 26, “a wavelength division multiplexer, the input end of the wavelength division multiplexer is in optical communication connection with the output end of the first modulator, the wavelength division multiplexer is used for decomposing the first modulation light signal into a plurality of wavelength division optical signals”, thus a wavelength division demultiplexer is disclosed because Wang teaches a wavelength division multiplexer that decomposes an optical signal into a plurality of wavelength division optical signals, thereby separating the signal based on wavelength, which corresponds to a wavelength division demultiplexer) coupled to the optical frequency comb source and configured to:
receive the third optical signal (Wang, Page 2 – line 23, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the third optical signals is disclosed because Wang teaches that the modulator is in optical communication connection with the output of the light source, such that the optical signals generated by the light source are provided to and received at the input of the modulator)
And decompose the third optical signal into a plurality of fourth optical signals of different wavelengths (Wang, Page 5 – line 7, “The wavelength division multiplexer 30 divides the first modulated light signal composed of a plurality of wavelengths to obtain a plurality of split-wave optical signals, for example, a plurality of split-wave optical signals may be represented by C1 to C9”, & Page 8 – line 23, “Therefore, light of different wavelengths in the output flat waveguide is diffracted and focused to different output channel waveguide position; after outputting the output channel waveguide, finishing the wavelength distribution, namely de-multiplexing function. The inverse process of this process, that is, if the signal light is reversely input, the multiplexing function is finished, the principle is the same”, thus decompose the third optical signal into a plurality of fourth optical signals of different wavelengths is disclosed because Wang teaches that the wavelength division multiplexer divides an optical signal composed of a plurality of wavelengths into a plurality of split-wave optical signals and further teaches that light of different wavelengths is separated and directed to different output channel waveguides, thereby performing wavelength-based decomposition into multiple optical signals of different wavelengths)
And a plurality of modulators (Wang, Page 3 – line 4, “the second modulator has a plurality of sub-modulators, the plurality of sub-modulators are corresponding to the plurality of output channels; the number of the sub-modulator and the wavelength number of the optical signal, the element number in the first matrix, the element number in the second matrix are equal”, thus a plurality of modulators is disclosed because Wang teaches a second modulator comprising a plurality of sub-modulators corresponding to multiple output channels) coupled to the wavelength division demultiplexer and configured to:
receive the fourth optical signals (Wang, Page 2 – line 23, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the fourth optical signals is disclosed because Wang teaches that the modulator is in optical communication connection with the output of the light source, such that the optical signals generated by the light source are provided to and received at the input of the modulator)
And respectively modulate the elements to the fourth optical signals to obtain the first optical signals (Wang, Page 2 – line 23, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 4 – line 37, “the number of wavelengths of the optical signal output by the multi-wavelength light source 10 according to the number of elements in the first matrix, the number of wavelengths in this example can be 9, then the wavelength corresponding to the elements in the first matrix is respectively represented as λ1 λ9, The first modulator 20 simultaneously loads each element of the first matrix X to 9 wavelengths in the form of light intensity, and the signal loading interval of x1 to x9 is Δ”, thus respectively modulate the elements to the fourth optical signals to obtain the first optical signals is disclosed because Wang teaches that the first modulator loads elements of the first matrix onto optical signals output from the light source to obtain first modulated light signals. Wang further teaches that each element is loaded onto a corresponding wavelength of the optical signals in the form of light intensity, thereby indicating that the elements are respectively modulated onto the optical signals to produce modulated optical signals)
Regarding Claim 10, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 6 as cited above and Wang further teaches:
a wavelength division multiplexer (Wang, Page 2 – line 35, “a wavelength division multiplexer, the input end of the wavelength division multiplexer is in optical communication connection with the output end of the delay line, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus a wavelength division multiplexer is disclosed because Wang teaches a wavelength division multiplexer configured to receive a plurality of optical signals and converge them into a combined wave light signal, thereby combining multiple optical signals into a single multi-wavelength optical signal) configured to:
receive the first optical signals (Wang, Page 2 – line 35, “a wavelength division multiplexer, the input end of the wavelength division multiplexer is in optical communication connection with the output end of the delay line, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus receive the first optical signals is disclosed because Wang teaches that the wavelength division multiplexer is in optical communication connection with the output of the delay line and receives the plurality of optical signals provided thereto for further processing)
And combine the first optical signals into a second group of the first optical signals (Wang, Page 2 – line 35, “a wavelength division multiplexer, the input end of the wavelength division multiplexer is in optical communication connection with the output end of the delay line, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus combine the first optical signals into a second group of the first optical signals is disclosed because Wang teaches that the wavelength division multiplexer converges a plurality of optical signals into a combined wave light signal, thereby combining multiple optical signals into a grouped optical signal)
[…] coupled to the wavelength division multiplexer and […], wherein each of the first groups comprises a portion of the second group (Wang, Page 2 – line 35, “a wavelength division multiplexer, the input end of the wavelength division multiplexer is in optical communication connection with the output end of the delay line, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus the wavelength division multiplexer wherein each of the first groups comprises a portion of the second group is disclosed because Wang teaches converging a plurality of second light modulation signals into a combined wave light signal, thereby forming a grouped optical signal that includes the plurality of signals)
Wang does not explicitly teach a beam splitter […] configured to decompose the second group into the first groups, […].
However, Arash further teaches:
And a beam splitter […] configured to decompose the second group into the first groups, […] (Arash, Page 6 – line 32, “The optical splitter may include a beam splitter, the beam splitter including at least one surface that transmits a predetermined proportion of the power of the light wave at the input port and reflects the remaining proportion of the power of the light wave at the input port”, & Page 10 – line 18, “In some embodiments, the laser unit may include: a laser source configured to generate light; and an optical power splitter configured to divide the light generated by the laser source into light outputs, wherein each light output has a substantial The same power” thus a beam splitter configured to decompose the second group into the first groups is disclosed because Arash teaches a beam splitter and optical power splitter that divide an input optical signal into multiple output light signals, thereby forming multiple groups of optical signals from the input signal. Further, incorporating Arash’s beam splitter supports efficient optical signal distribution, as dividing a signal into multiple paths enables parallel processing while maintaining low power consumption)
Regarding Claim 11, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 6 as cited above and Arash further teaches:
wherein the optical splitting device comprises a planar waveguide configured to split, in space, the first optical signals into the first groups (Arash, Page 21 – line 19, “In some embodiments, the optical splitter may include a waveguide optical splitter that transmits a predetermined proportion of the power of the light wave guided by the input optical waveguide to the first output optical waveguide, and transfers the light wave guided by the input optical waveguide The remaining proportion of the power is sent to the second output optical waveguide. In some embodiments, the guided mode of the input optical waveguide may be adiabatically coupled to the complex guided mode of each of the first output optical waveguide and the second output optical waveguide”, & Page 61 – line 31, “The above describes the use of photonic circuits including Mach-Zehnder interferometers, directional couplers, planar optical waveguides, and photodetectors to implement logic gates (such as AND, OR, and XOR gates)”, thus wherein the optical splitting device comprises a planar waveguide configured to split, in space, the first optical signals into the first groups is disclosed because Arash teaches a waveguide optical splitter that divides an input optical signal into multiple output optical waveguides, thereby spatially splitting the optical signals into multiple groups, and further teaches planar optical waveguides as part of the optical circuit)
Regarding Claim 12, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 6 as cited above and Arash further teaches:
an optical splitting element configured to split the first optical signals of different wavelengths into n groups of the first optical signals, wherein n is a quantity of elements in the second data and is an integer greater than one (Arash, Page 5 – line 35, “the first set of optical amplitude modulators are used to encode the corresponding complex optical signals carried by the optical waveguides A set of multiple input values. The computing device includes a plurality of replication modules, and for each of at least two subsets of one or more optical signals, a corresponding set of one or more of the replication modules is configured to combine the one or more optical signals The subset is divided into two or more copies of the optical signal”, & Page 6 – line 26, “The optical splitter may include a waveguide optical splitter, which transmits a predetermined ratio of the power of the light wave guided by the input optical waveguide of the replication module to the first output optical waveguide of the replication module, and transfers it from the input of the replication module. The remaining proportion of the power of the light wave guided by the optical waveguide is sent to the second output optical waveguide of the replication module”, thus an optical splitting element configured to split the first optical signals of different wavelengths into n groups of the first optical signals is disclosed because Arash teaches that replication modules divide optical signals into two or more copies, thereby forming multiple groups of the optical signals, and further teaches a waveguide optical splitter that splits an input optical signal into multiple output optical waveguides, thereby physically implementing the splitting into multiple groups)
And a planar waveguide coupled to the optical splitting element and configured to transmit the n groups of the first optical signals to the modulator array (Arash, Page 6 – line 26, “The optical splitter may include a waveguide optical splitter, which transmits a predetermined ratio of the power of the light wave guided by the input optical waveguide of the replication module to the first output optical waveguide of the replication module, and transfers it from the input of the replication module. The remaining proportion of the power of the light wave guided by the optical waveguide is sent to the second output optical waveguide of the replication module” & Page 61 – line 31, “The above describes the use of photonic circuits including Mach-Zehnder interferometers, directional couplers, planar optical waveguides, and photodetectors to implement logic gates (such as AND, OR, and XOR gates)”, thus a planar waveguide coupled to the optical splitting element and configured to transmit the n groups of the first optical signals to the modulator array is disclosed because Arash teaches a waveguide optical splitter that directs split optical signals from an input optical waveguide into multiple output optical waveguides, thereby transmitting the split optical signals, and further teaches planar optical waveguides as part of the photonic circuit, which are used to route optical signals between components including modulators),
wherein the modulator array comprises n modulators (Arash, Page 36 – line 2, “The modulator array 144 is configured to receive light input from the laser unit 142, and modulate the intensity of the received light input based on the modulator control signal (which is an electrical signal). Examples of modulators include Mach-Zehnder interference (MZI) modulators, ring resonator modulators, and electro-absorption modulators. The modulator array 144 has N modulators, and each modulator receives one of the N light outputs of the laser unit 142. The modulator receives the control signal corresponding to the element of the digital input vector and modulates the intensity of the light. The control signal can be generated by the DAC unit 130”, thus wherein the modulator array comprises n modulators is disclosed because Arash teaches that the modulator array has N modulators, each configured to receive a corresponding light output, thereby indicating a modulator array comprising a plurality of modulators equal in number to the optical signals),
wherein each of the modulators is configured to receive one of the n groups (Arash, Page 36 – line 1, “The modulator array 144 is configured to receive light input from the laser unit 142, and modulate the intensity of the received light input based on the modulator control signal (which is an electrical signal). Examples of modulators include Mach-Zehnder interference (MZI) modulators, ring resonator modulators, and electro-absorption modulators. The modulator array 144 has N modulators, and each modulator receives one of the N light outputs of the laser unit 142. The modulator receives the control signal corresponding to the element of the digital input vector and modulates the intensity of the light. The control signal can be generated by the DAC unit 130”, & Page 5 – line 36, “The computing device includes a plurality of replication modules, and for each of at least two subsets of one or more optical signals, a corresponding set of one or more of the replication modules is configured to combine the one or more optical signals The subset is divided into two or more copies of the optical signal”, thus wherein each of the modulators is configured to receive one of the n groups is disclosed because Arash teaches that the modulator array includes N modulators, each receiving a corresponding one of N optical signals, thereby establishing a one-to-one correspondence between modulators and optical signals, and further teaches that optical signals are divided into two or more copies, thereby forming multiple groups of optical signals such that each modulator receives a respective group),
and wherein each of the n groups comprises the first optical signals (Arash, Page 5 – line 36, “The computing device includes a plurality of replication modules, and for each of at least two subsets of one or more optical signals, a corresponding set of one or more of the replication modules is configured to combine the one or more optical signals The subset is divided into two or more copies of the optical signal”, thus wherein each of the n groups comprises the first optical signals is disclosed because Arash teaches that optical signals are divided into two or more copies using replication modules, such that each resulting group comprises optical signals from the same original set)
Regarding Claim 13, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Wang further teaches:
wherein the modulator array comprises a plurality of modulators (Wang, Page 3 – line 4, “the second modulator has a plurality of sub-modulators, the plurality of sub-modulators are corresponding to the plurality of output channels; the number of the sub-modulator and the wavelength number of the optical signal, the element number in the first matrix, the element number in the second matrix are equal”, thus a plurality of modulators is disclosed because Wang teaches a second modulator comprising a plurality of sub-modulators corresponding to multiple output channels), and wherein each of the modulators is configured to:
And output a plurality of second intermediate optical signals based on the modulated second data (Wang, Page 2 – line 44, “the second modulator is further configured to correspondingly load a plurality of elements in the second matrix to be processed to the plurality of divided-wave optical signals to obtain a plurality of second modulated optical signals”, & Page 5 – line 22, “the second modulator 40 can be a plurality of elements in the second matrix Y in the form of light intensity is loaded on the plurality of wavelength division optical signal, finally forming a plurality of second modulated light signal. As shown in FIG. 4, the y1 is loaded to C1, then obtaining x1y1, x2y1 ... x9y1, corresponding y2 is loaded to C2, then obtaining x1y2, x2y2 ... x9y2, and so on y9 is loaded to C9, then obtaining x1y9, x2y9 ... x9y9”, thus output a plurality of second intermediate optical signals based on the modulated second data is disclosed because Wang teaches that the second modulator loads elements of the second matrix onto wavelength division optical signals to generate a plurality of second modulated optical signals, including products of elements from the first and second data, thereby producing multiple optical signals based on the modulated second data)
Regarding Claim 14, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Wang further teaches:
wherein the computing result comprises a result of a convolution operation on the first data and the second data (Wang, Page 5-6 – line 39, “the input end of the photoelectric detector 70 is electrically connected with the output end of the wavelength division multiplexer 60, for sampling the combined wave light signal and converting into electric signal, outputting the convolution result”, & Page 2 – line 3, “The convolutional neural network is a neural network model widely used in deep learning, especially has important application in the research field of machine vision, image recognition and so on. the convolution kernel in the convolutional neural network is a basic operation unit of image blurring, sharpening, edge detection and other processing schemes, generally under the condition of a small-scale matrix. using the convolution kernel to slide on the large scale matrix of the image information; multiplying the pixel grey value on the image point with the numerical value on the corresponding convolution kernel, then adding; re-weighting the information of each pixel of the image, and realizing image feature extraction”, thus wherein the computing result comprises a result of a convolution operation on the first data and the second data is disclosed because Wang teaches outputting a convolution result from the photoelectric detector based on optical signals formed from multiplication of elements of input data and a convolution kernel, thereby representing a convolution operation on the first data and the second data)
Regarding Claim 15, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Wang further teaches:
wherein the optical computing apparatus comprises a chip (Wang, Page 6 – line 17, “In addition, the optical signal itself is an analog signal; theoretically it can perform high precision floating point operation. floating point data can be converted into light intensity; the conversion precision is related to the linearity of the optoelectronic device, receiving end sensitivity and system noise. Compared with using electronic as information carrier for operation, the on-chip optical computing architecture can avoid the process of binary conversion of data, realizing high speed floating point operation.”, thus wherein the optical computing apparatus comprises a chip is disclosed because Wang teaches an on-chip optical computing architecture, thereby indicating that the optical computing apparatus is implemented on a chip)
Regarding Claim 16,
an optical computing system (Wang, Claim 1, “A photonic convolutional neural network system, wherein it comprises: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths”, thus an optical computing system is disclosed because Wang teaches a photonic convolutional neural network system comprising optical components configured to perform optical signal processing) comprising:
an optical computing apparatus (Wang, Claim 1, “A photonic convolutional neural network system, wherein it comprises: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths;”, thus the disclosed photonic convolutional neural network system comprises interconnected optical components that generate, modulate, and process optical signals to perform convolution operations, and therefore constitutes an optical computing apparatus) comprising:
a light source array configured to send a plurality of first groups of first optical signals based on first data to be computed, wherein each of the first groups represents a plurality of elements in the first data (Wang, Page 2 – line 21, “The embodiment of the present invention provides a photonic convolutional neural network system, comprising: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths; a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 3 – line 1, “The photonic convolutional neural network system according to one embodiment of the present invention, the output end of the wavelength division multiplexer comprises a plurality of output channels, each of the output channel corresponding to output one of the wavelength division optical signal; the second modulator has a plurality of sub-modulators, the plurality of sub-modulators are corresponding to the plurality of output channels”, thus a light source array configured to send a plurality of first groups of first optical signals based on first data to be computed, wherein each of the first groups represents a plurality of elements in the first data is disclosed because Wang teaches a multi-wavelength light source that outputs optical signals having a plurality of wavelengths, corresponding to a plurality of optical signals. Wang further teaches that a first modulator loads elements of a first matrix onto the optical signals, such that the optical signals carry the first data, and that the signals are output via multiple channels. Accordingly, the multi-wavelength light source corresponds to the light source array, the plurality of wavelengths corresponds to the plurality of first optical signals, and the loaded elements correspond to the plurality of elements represented by each group)
[…] coupled to the light source array and configured to:
receive the groups (Wang, Page 2- line 21, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the groups is disclosed because Wang teaches that the first modulator is in optical communication connection with the multi-wavelength light source and receives the optical signals output therefrom. Since the optical signals include multiple wavelengths carrying elements of the first matrix / the first data, the first modulator necessarily receives the groups of optical signals representing the elements)
modulate, based on second data, the groups to provide modulated second data (Wang, Page 2- line 21, “The embodiment of the present invention provides a photonic convolutional neural network system, comprising: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths; a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 2 – line 29, “a second modulator, the input end of the second modulator is in optical communication connection with the output end of the wavelength division multiplexer, the second modulator is used for loading the element in the second matrix to be processed to the plurality of wavelength division optical signals to obtain a plurality of second modulation optical signals”, thus modulate, based on second data, the groups to provide modulated second data is disclosed because Wang teaches that the first modulator loads elements of a first matrix onto optical signals to obtain a first modulated light signal, and that a second modulator further loads elements of a second matrix onto the optical signals to obtain second modulated optical signals. Since the optical signals carrying the first data are further modulated with additional data, this corresponds to modulating the first groups to produce modulated second data)
and output a plurality of intermediate optical signals based on the modulated second data (Wang, Page 5 – line 14, “For example, the second modulator is used for respectively loading a plurality of elements in the second matrix to be processed to a plurality of sub-wave optical signals to obtain a plurality of second modulated light signals”, & Page 5 – line 25, “As shown in FIG. 4, the y1 is loaded to C1, then obtaining x1y1, x2y1 ... x9y1, corresponding y2 is loaded to C2, then obtaining x1y2, x2y2 ... x9y2, and so on y9 is loaded to C9, then obtaining x1y9, x2y9 ... x9y9”, thus output a plurality of intermediate optical signals based on the modulated second data is disclosed because Wang teaches that the second modulator loads elements of the second matrix onto the optical signals to generate a plurality of second modulated light signals. Wang further discloses that this modulation produces multiple outputs (e.g., x1y1, x2y1 … x9y9), which are optical signals resulting from the applied second data. The plurality of second modulated light signals corresponds to a plurality of intermediate optical signals generated based on the modulated second data)
And a wavelength router (Wang, Page 8 – line 12, “Regarding the on-chip planar array waveguide grating, AWG (Arrayed Waveguide Grating) is the preferred technology in dense wavelength division multiplexing system (DWDM). A group of grating formed by array waveguide with equal length difference, using the ability of dividing wave”, Page 8 – line 23, “Therefore, light of different wavelengths in the output flat waveguide is diffracted and focused to different output channel waveguide position; after outputting the output channel waveguide, finishing the wavelength distribution, namely de-multiplexing function. The inverse process of this process, that is, if the signal light is reversely input, the multiplexing function is finished, the principle is the same”, thus a wavelength router is disclosed because Wang teaches an arrayed waveguide grating (AWG) that separates optical signals by wavelength and directs them to different output channel waveguides. Wang also discloses that light of different wavelengths is diffracted and focused to different output channel positions, indicating routing of wavelengths to distinct paths) comprising […] wherein the wavelength router is configured to:
receive the intermediate optical signals […] (Wang, Page 5 – line 36, “the input end of the wavelength division multiplexer 60 is in optical communication connection with the output end of the delay line 50, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus receive the intermediate optical signals is disclosed because Wang teaches that the wavelength division multiplexer is in optical communication connection with the output of the delay line and receives the plurality of second modulated light signals. Since these second modulated light signals correspond to intermediate optical signals, the wavelength division multiplexer receives the intermediate optical signals)
route, based on […] a wavelength of each intermediate optical signal […], the intermediate optical signals as a plurality of second optical signals […], wherein intermediate optical signals corresponding to products of elements of the first data and the second data […] (Wang, Page 5, “the input end of the second modulator 40 is connected with the output end of the wavelength division multiplexer 30 optical communication connection, the second modulator 40 for the element to be processed in the second matrix is loaded to a plurality of wavelength division optical signal to obtain a plurality of second modulated light signal”, & Page 7, “the y1, y2 ... y9 are loaded to C1, then obtaining the x1y1, x1y2 ... x1y9, the y1, y2 ... y9 are loaded to C2, then obtaining the x2y1, x2y2 ... x2y9, and so on, y2 ... y9 are loaded to C9, then obtaining x9y1, x9y2 ... x9y9”, & Page 8, “for different wavelengths of light, the phase difference is different, Therefore, light of different wavelengths in the output flat waveguide is diffracted and focused to different output channel waveguide position; after outputting the output channel waveguide, finishing the wavelength distribution, namely de-multiplexing function”, thus route, based on […] a wavelength of each intermediate optical signal […], the intermediate optical signals as a plurality of second optical signals […], wherein intermediate optical signals corresponding to products of elements of the first data and the second data […] is disclosed because Wang teaches that wavelength-division optical signals
C
1
through
C
9
carry elements
x
1
through
x
9
of the first matrix, corresponding to the first data. Wang further teaches loading elements
y
1
through
y
9
of the second matrix, corresponding to the second data, onto the wavelength-division optical signals to produce second-modulated optical signals having values
x
i
y
j
. The second-modulated optical signals therefore correspond to the intermediate optical signals representing products of elements of the first data and the second data. Wang also teaches that light of different wavelengths is diffracted and focused to different output-channel waveguide positions, such that the wavelengths of the intermediate optical signals are used to route the intermediate optical signals as a plurality of second optical signals)
and output the second optical signals to indicate a computing result of the first data and the second data (Wang, Page 9 – claim 1, “a photoelectric detector, the input end of the photoelectric detector is in optical communication connection with the output end of the wavelength division multiplexer, for sampling the combined wave light signal and converting into electric signal, outputting the convolution result”, & Page 6 – line 2, “It can be understood that, as shown in FIG. 5, the photoelectric detector 70 can select the proper sampling time, sampling the information carried on the combined wave optical signal, for example, according to the sampling time of the obtained first matrix X and the second matrix Y by bit multiplying result, namely only collecting the x1y1 in the wave-combining optical signal. x2y2 ... x9y9 information, and converting it into electric signal, finally obtaining the convolution result”, thus output the second optical signals, wherein the second optical signals to indicate a computing result of the first data and the second data is disclosed because Wang teaches that the photoelectric detector samples the combined optical signal carrying results of operations between elements of the first matrix and the second matrix (e.g., x1y1, x2y2 … x9y9) and outputs a convolution result. Since these optical signals encode the results of the computation between the first and second data, they correspond to second optical signals indicating a computing result)
a processor (Wang, Page 3 – line 13, “a processor, the input end of the processor is electrically connected with the output end of the photoelectric detector, for compensating the convolution result”, thus a processor is disclosed) coupled to the optical computing apparatus and configured to send at least one of the first data or the second data to the optical computing apparatus.
Wang does not explicitly teach a modulator array, a plurality of input ports coupled to the modulator array and a plurality of output ports, receiving at the plurality of input ports, and […] to the plurality of output ports[…], […] based on both a wavelength […] and an identity of one of the plurality of input ports at which […] is received […]are routed to the same output port, […], […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data, and […] with an equal sequence number difference (i-j) […].
However, Arash teaches:
a modulator array (Arash, Page 36 – line 1, “The modulator array 144 is configured to receive light input from the laser unit 142, and modulate the intensity of the received light input based on the modulator control signal (which is an electrical signal). Examples of modulators include Mach-Zehnder interference (MZI) modulators, ring resonator modulators, and electro-absorption modulators”, thus a modulator array is disclosed because Arash teaches a modulator array configured to receive light input and modulate the intensity of the received light based on control signals)
a plurality of input ports coupled to the modulator array and a plurality of output ports (Arash, Page 10 – line 25, “an input waveguide array for receiving a light input vector; an optical interference unit, which is in optical communication with the input waveguide array, for performing a conversion of the light input vector into a second optical signal array Linear conversion; and the output waveguide array, in optical communication with the optical interference unit, for guiding the second optical signal array, wherein at least one input waveguide in the input waveguide array passes through the optical interference unit and each output waveguide in the output waveguide array Optical communications”, thus a plurality of input ports coupled to the modulator array and a plurality of output ports is disclosed because Arash teaches an input waveguide array for receiving light input and an output waveguide array for guiding optical signals. Since waveguides function as optical input and output interfaces, the input waveguide array corresponds to a plurality of input ports and the output waveguide array corresponds to a plurality of output ports)
[…] to the plurality of output ports[…] (Arash, Page 40-41, “The OMM unit 150 may include an array of input waveguides 152 to receive light input vectors; an optical interference unit 154 in optical communication with the array of input waveguides 152; and an array of output waveguides 156 in optical communication with the optical interference unit 154. The optical interference unit 154 linearly converts the light input vector into a second optical signal array. The array of output waveguides 156 guides the second array of optical signals output by the optical interference unit 154. At least one input waveguide in the array of input waveguides 152 optically communicates with each output waveguide in the array of output waveguides 156 through the optical interference unit 154. For example, for an optical input vector of length N, the OMM unit 150 may include N input waveguides 152 and N output waveguides 156’, thus […] to the plurality of output ports […] is disclosed because Arash teaches that the optical interference unit converts the optical input vector into a second array of optical signals and that an array of
N
output waveguides guides those signals. The N output waveguides correspond to the plurality of output ports, and guiding the second array of optical signals through the output waveguides corresponds to directing the second optical signals to the plurality of output ports
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Wang’s approach of sending a plurality of optical signals using a light source array and routing the optical signals based on wavelengths using a wavelength router to perform computation with Arash’s approach of using a modulator array coupled to a plurality of input ports and a plurality of output ports for receiving at the plurality of input ports and routing to the plurality of output ports, thereby reducing power consumption by leveraging the modulator array architecture optimized for efficient optical signal modulation (Arash, Page 37 – line 12, “In some embodiments, the modulator of the modulator array 144 and/or the OMM unit 150 can be designed to reduce power consumption, so that when the modulator is operated to generate a modulation value representing a more frequently occurring coefficient It consumes less power while operating, and consumes more power when operating the modulator to generate modulation values representing coefficients that occur less frequently. For example, for certain data sets that are known to have certain characteristics, power consumption can be reduced. Figure 42 shows the modulation value probability distribution diagram 4200 (dotted line) superimposed on the modulator power diagram 4202 (solid line) for the specific design of the modulator and/or OMM unit 150 of the modulator array 144”, & Page 38 – line 4, “The power distribution shown in Figure 42 shows that the zero modulation power is used to achieve the zero modulation value, but in other embodiments, there may be residual low but non-zero modulation power at the zero modulation value. For these low coefficient weighted data sets, the power consumption can usually be reduced by using a modulator”, thereby reducing power consumption by leveraging the modulator array optimized for efficient optical signal modulation is disclosed because Arash teaches that the modulator array can be designed such that modulation values corresponding to more frequently occurring coefficients consume less power, resulting in an overall reduction in power consumption for optical signal processing)
Wang combined with Arash does not explicitly teach […] based on both a wavelength […] and an identity of one of the plurality of input ports at which […] is received […]are routed to the same output port, […], […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data, and […] with an equal sequence number difference (i-j) […].
However, Roberto teaches:
[…]based on both a wavelength […] and an identity of one of the plurality of input ports at which […] is received […]are routed to the same output port, […] (Roberto, Page 1 – Section 1, “An arrayed waveguide grating router (AWGR) [7], [8] is an example of devices with such wavelength routing capability. As Fig. 1(a) and (b) illustrate, the well-known wavelength routing property of an AWGR allows any input port to communicate with any output port simultaneously using different wavelengths without contention”, & Page 2 – Fig 1 a and b,
PNG
media_image1.png
412
489
media_image1.png
Greyscale
,
& Page 8 – Section V, “Second, we show a scenario with network contention where C1 and C2 send packets to C3 with random arrival time. To avoid resynchronization in the FPGA GTH receiver, tunable optical delay lines in each transmission path guarantee clock-phase matching of the packets at AWGR output 3”, thus […] based on both […] a wavelength […] and an identity of one of the plurality of input ports at which […] is received […] are routed to the same output port […] is disclosed because Roberto teaches an AWGR in which the destination output port is determined by the combination of the wavelength used and the input port through which the signal enters, as illustrated by the wavelength-routing property and wavelength-assignment table of Figures 1(a) and 1(b). Roberto further teaches that packets originating from C1 and C2 are directed to the same AWGR output 3)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Wang and Arash with Roberto by incorporating Roberto’s AWGR wavelength-routing technique into the optical computing system of Wang and Arash. Wang teaches generating wavelength division optical signals that represent products of elements of first and second matrices, while Arash teaches communicating optical signals through arrays of input and output waveguides. Roberto teaches that an AWGR routes signals between input and output ports based on wavelength and permits simultaneous optical communication without contention. Therefore, a POSITA would have been motivated to use Roberto’s AWGR to route Wang’s wavelength encoded product signals from Arash’s input waveguides to selected output waveguides based on the wavelength of each signal and the input port through which the signal is received. This would allow multiple product signals to be routed in parallel, reduce contention, and improve the speed, scalability, and energy efficiency of the optical convolution system (Roberto, Page 1 - Section I, “Compared to electrical interconnects, optical interconnects provide (1) higher transmission bandwidth and lower energy consumption independently of distance, (2) inherent parallelism, (3) low interference and crosstalk, and (4) low parasitic” and “the well-known wavelength routing property of an AWGR allows any input port to communicate with any output port simultaneously using different wavelengths without contention”)
Wang and Arash combined with Roberto does not explicitly teach […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data, and […] with an equal sequence number difference (i-j) […].
However, Xu teaches:
[…] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data (Xu, Page 3, “The input data vector X is encoded as the intensity of temporal symbols in a serial electrical waveform at a symbol rate 1/τ (baud), where τ is the symbol period. The convolution kernel is similarly represented by a weight vector W of length R that is then encoded in the optical power of the microcomb lines through spectral shaping performed by a Waveshaper. The temporal waveform X is then multi cast onto the kernel wavelength channels via electro-optical modulation, thus generating the replicas weighted by W. Next the optical waveform is transmitted through a dispersive delay with a delay step (between adjacent wavelength channels) equal to the symbol duration of X, effectively achieving time and wavelength interleaving. Finally, the delayed and weighted replicas are summed via high speed photodetection so that each time slot yields a convolution between X and W for a given convolution window, or receptive field. As such, the convolution window effectively slides at the modulation speed matching the baud rate of X. Each output symbol is the result of R multiply-and-accumulate operations”, thus […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data is disclosed because Xu teaches that the input vector X, corresponding to the first data, is encoded using optical intensity and that the convolution kernel W, corresponding to the second data, is encoded using the optical power of respective wavelength channels. Xu further teaches generating weighted optical replicas representing products of elements of X and W, aligning and summing those replicas through photodetection, and producing an output symbol for each time slot. Because each output symbol results from the multiply-and-accumulate operations for one convolution window, each output symbol represents one element of the convolution result. In the combined system, applying this intensity-based convolution technique to the plurality of output ports results in the light intensity at each output port representing a respective element of the convolution result)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to further combine Wang, Arash, and Roberto with Xu by incorporating Xu’s intensity based optical convolution and photodetection technique into the combined optical routing system. Wang teaches producing optical signals corresponding to products of elements of two matrices, Arash teaches guiding optical signals through a plurality of output waveguides, and Roberto teaches routing the optical signals to selected output ports. Xu teaches encoding input data and convolution weights using optical intensity and optical power, generating weighted optical replicas, and summing those replicas through high-speed photodetection so that each output symbol represents the multiply-and-accumulate result for a convolution window. Therefore, a POSITA would have been motivated to apply Xu’s photodetection technique to the product signals routed to each output port so that the signals received at that port are summed and represented as a respective convolution-output element. This would provide a direct and high speed technique for obtaining convolution results from the routed optical product signals and would take advantage of the parallel wavelength channels already used by the combined system (Xu, Page 3, “Finally, the delayed and weighted replicas are summed via high speed photodetection so that each time slot yields a convolution between X and W for a given convolution window, or receptive field. As such, the convolution window effectively slides at the modulation speed matching the baud rate of X. Each output symbol is the result of R multiply-and-accumulate operations , with the computing speed given by 2R/τ FLOPS. Since the speed of this process scales with both the baud rate and number of wavelengths, it can be dramatically boosted into the Tera-FLOP regime by using the massively parallel wavelength channels of the microcomb source”)
Wang, Arash, and Roberto combined with Xu does not explicitly teach […] with an equal sequence number difference (i-j) […].
However, Brea teaches:
[…] with an equal sequence number difference (i-j) […] (Brea, Par. [0192], “Each output vector sample y.sub.p is produced by the time reversal of the matrix sequence {C.sub.p,p=0,1,2, . . . ,M-1} and its circular shifting to the right. The resulting operation shall be called circular matrix convolution. Note that the sum of the indices of each sample product inside the summation modulo M is equal to the index of the output being calculated by the circular matrix convolution”, thus Brea teaches grouping product terms according to a common convolution-output index. Because one of the sequences is time-reversed before the products are formed, the sum-of-indices relationship for the reversed sequence corresponds to a difference
i
j
between the original sequence numbers. Product terms having the same sequence-number difference therefore contribute to the same convolution-output element)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to further combine Wang, Arash, Roberto, and Xu with Brea by using Brea’s known convolution indexing relationship to determine which optical product signals should be routed to the same output port. Wang teaches optical signals representing products of first data and second-data elements, Roberto teaches routing optical signals according to wavelength and input-port identity, and Xu teaches summing product signals to produce individual convolution-output elements. Brea teaches that the product terms contributing to a convolution output are grouped according to a common output index and that one sequence is time reversed before the product terms are formed. Therefore, a POSITA would have been motivated to assign the wavelengths and input ports of the product signals according to Brea’s convolution index relationship so that products having the same sequence number difference (i-j) are directed to the same output port. Xu’s photodetection technique could then sum the grouped product signals at that port to produce the corresponding convolution element. This would implement the known mathematical organization of convolution products directly in the optical routing hardware and allow multiple convolution elements to be calculated efficiently and in parallel (Brea, Paragraph [0192], “Each output vector sample y.sub.p is produced by the time reversal of the matrix sequence {C.sub.p, p=0,1,2, . . . ,M-1} and its circular shifting to the right. The resulting operation shall be called circular matrix convolution. Note that the sum of the indices of each sample product inside the summation modulo M is equal to the index of the output being calculated by the circular matrix convolution”)
Regarding Claim 17, Wang teaches a convolution computing method implemented by an optical computing apparatus, wherein the convolution computing method comprises:
sending, by a light source array in the optical computing apparatus, a plurality of groups of first optical signals based on first data to be computed, wherein each of the groups represents a plurality of elements in the first data (Wang, Page 2 – line 21, “The embodiment of the present invention provides a photonic convolutional neural network system, comprising: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths; a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 3 – line 1, “The photonic convolutional neural network system according to one embodiment of the present invention, the output end of the wavelength division multiplexer comprises a plurality of output channels, each of the output channel corresponding to output one of the wavelength division optical signal; the second modulator has a plurality of sub-modulators, the plurality of sub-modulators are corresponding to the plurality of output channels”, thus sending, by a light source array in the optical computing apparatus, a plurality of groups of first optical signals based on first data to be computed, wherein each of the groups represents a plurality of elements in the first data is disclosed because Wang teaches a multi-wavelength light source that outputs optical signals which are modulated with elements of the first matrix to form first optical signals, and further teaches that the optical signals are divided into multiple wavelength channels and output channels, thereby forming multiple groups of optical signals, each carrying information corresponding to elements of the first data)
receiving […] the groups (Wang, Page 2- line 21, “a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, thus receive the groups is disclosed because Wang teaches that the first modulator is in optical communication connection with the multi-wavelength light source and receives the optical signals output therefrom. Since the optical signals include multiple wavelengths carrying elements of the first matrix / the first data, the first modulator necessarily receives the groups of optical signals representing the elements)
modulating, based on second data, the groups to provide modulated second data (Wang, Page 2- line 21, “The embodiment of the present invention provides a photonic convolutional neural network system, comprising: a multi-wavelength light source for outputting an optical signal having a plurality of wavelengths; a first modulator, the input end of the first modulator is in optical communication connection with the output end of the multi-wavelength light source, the first modulator is used for loading the element in the first matrix to be processed to the optical signal to obtain the first modulated light signal”, & Page 2 – line 29, “a second modulator, the input end of the second modulator is in optical communication connection with the output end of the wavelength division multiplexer, the second modulator is used for loading the element in the second matrix to be processed to the plurality of wavelength division optical signals to obtain a plurality of second modulation optical signals”, thus modulating the groups to provide modulated second data is disclosed because Wang teaches that the first modulator loads elements of a first matrix onto optical signals to obtain a first modulated light signal, and that a second modulator further loads elements of a second matrix onto the optical signals to obtain second modulated optical signals. Since the optical signals carrying the first data are further modulated with additional data, this corresponds to modulating the first groups to produce modulated second data)
outputting […] a plurality of intermediate optical signals based the modulated second data (Wang, Page 5 – line 14, “For example, the second modulator is used for respectively loading a plurality of elements in the second matrix to be processed to a plurality of sub-wave optical signals to obtain a plurality of second modulated light signals”, & Page 5 – line 25, “As shown in FIG. 4, the y1 is loaded to C1, then obtaining x1y1, x2y1 ... x9y1, corresponding y2 is loaded to C2, then obtaining x1y2, x2y2 ... x9y2, and so on y9 is loaded to C9, then obtaining x1y9, x2y9 ... x9y9”, thus outputting a plurality of intermediate optical signals based the modulated second data is disclosed because Wang teaches that the second modulator loads elements of the second matrix onto a plurality of sub-wave optical signals corresponding to different channels (groups) to generate a plurality of second modulated optical signals, including products of elements of the first data and the second data, thereby producing intermediate optical signals based on the groups and the modulated second data)
receiving […] the intermediate optical signals (Wang, Page 5 – line 36, “the input end of the wavelength division multiplexer 60 is in optical communication connection with the output end of the delay line 50, for converging the plurality of second light modulation signals after being rearranged to output as the combined wave light signal”, thus receiving the intermediate optical signals is disclosed because Wang teaches that the wavelength division multiplexer is in optical communication connection with the output of the delay line and receives the plurality of second modulated light signals. Since these second modulated light signals correspond to intermediate optical signals, the wavelength division multiplexer receives the intermediate optical signals)
routing, […] the intermediate optical signals as a plurality of second optical signals […], wherein intermediate optical signals corresponding to products of elements of the first data and the second data […] (Wang, Page 5, “the input end of the second modulator 40 is connected with the output end of the wavelength division multiplexer 30 optical communication connection, the second modulator 40 for the element to be processed in the second matrix is loaded to a plurality of wavelength division optical signal to obtain a plurality of second modulated light signal”, & Page 7, “the y1, y2 ... y9 are loaded to C1, then obtaining the x1y1, x1y2 ... x1y9, the y1, y2 ... y9 are loaded to C2, then obtaining the x2y1, x2y2 ... x2y9, and so on, y2 ... y9 are loaded to C9, then obtaining x9y1, x9y2 ... x9y9”, & Page 8, “for different wavelengths of light, the phase difference is different, Therefore, light of different wavelengths in the output flat waveguide is diffracted and focused to different output channel waveguide position; after outputting the output channel waveguide, finishing the wavelength distribution, namely de-multiplexing function”, thus routing, […] the intermediate optical signals as a plurality of second optical signals […], wherein intermediate optical signals corresponding to products of elements of the first data and the second data […] is disclosed because Wang teaches that wavelength-division optical signals
C
1
through
C
9
carry elements
x
1
through
x
9
of the first matrix, corresponding to the first data. Wang further teaches loading elements
y
1
through
y
9
of the second matrix, corresponding to the second data, onto the wavelength-division optical signals to produce second-modulated optical signals having values
x
i
y
j
. The second-modulated optical signals therefore correspond to the intermediate optical signals representing products of elements of the first data and the second data. Wang also teaches that light of different wavelengths is diffracted and focused to different output-channel waveguide positions, such that the wavelengths of the intermediate optical signals are used to route the intermediate optical signals as a plurality of second optical signals)
and outputting the second optical signals to indicate a computing result of the first data and the second data (Wang, Page 9 – claim 1, “a photoelectric detector, the input end of the photoelectric detector is in optical communication connection with the output end of the wavelength division multiplexer, for sampling the combined wave light signal and converting into electric signal, outputting the convolution result”, & Page 6 – line 2, “It can be understood that, as shown in FIG. 5, the photoelectric detector 70 can select the proper sampling time, sampling the information carried on the combined wave optical signal, for example, according to the sampling time of the obtained first matrix X and the second matrix Y by bit multiplying result, namely only collecting the x1y1 in the wave-combining optical signal. x2y2 ... x9y9 information, and converting it into electric signal, finally obtaining the convolution result”, outputting the second optical signals to indicate a computing result of the first data and the second data is disclosed because Wang teaches that second modulated optical signals representing products of elements of the first and second data are combined and provided to a photoelectric detector, which samples the combined optical signal and outputs a convolution result, thereby producing a computing result based on the second optical signals that represents a computation on the first data and the second data)
Wang does not explicitly disclose a modulator array, a plurality of input ports coupled to the modulator array and a plurality of output ports, receiving at the plurality of input ports, […] to a plurality of output ports of the wavelength router […], […] based on both a wavelength of each intermediate optical signal and an identity of one of the plurality of input ports at which the intermediate optical signal is received […] are routed to the same output port […], […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data […], and […] with an equal sequence number difference (i-j) […].
However, Arash teaches:
[…], by a modulator array in the optical computing apparatus, […] (Arash, Page 36 – line 1, “The modulator array 144 is configured to receive light input from the laser unit 142, and modulate the intensity of the received light input based on the modulator control signal (which is an electrical signal). Examples of modulators include Mach-Zehnder interference (MZI) modulators, ring resonator modulators, and electro-absorption modulators”, thus a modulator array is disclosed because Arash teaches a modulator array configured to receive light input and modulate the intensity of the received light based on control signals)
[…], by a plurality of input ports of a wavelength router in the optical computing apparatus, […] (Arash, Page 10 – line 25, “an input waveguide array for receiving a light input vector; an optical interference unit, which is in optical communication with the input waveguide array, for performing a conversion of the light input vector into a second optical signal array Linear conversion; and the output waveguide array, in optical communication with the optical interference unit, for guiding the second optical signal array, wherein at least one input waveguide in the input waveguide array passes through the optical interference unit and each output waveguide in the output waveguide array Optical communications”, thus receiving by a plurality of input ports of a wavelength router in the optical computing apparatus is disclosed because Arash teaches an input waveguide array configured to receive optical signals, wherein the input waveguides serve as input ports for receiving optical signal inputs and routing them through the optical system)
[…] to a plurality of output ports of the wavelength router […] (Arash, Page 40-41, “The OMM unit 150 may include an array of input waveguides 152 to receive light input vectors; an optical interference unit 154 in optical communication with the array of input waveguides 152; and an array of output waveguides 156 in optical communication with the optical interference unit 154. The optical interference unit 154 linearly converts the light input vector into a second optical signal array. The array of output waveguides 156 guides the second array of optical signals output by the optical interference unit 154. At least one input waveguide in the array of input waveguides 152 optically communicates with each output waveguide in the array of output waveguides 156 through the optical interference unit 154. For example, for an optical input vector of length N, the OMM unit 150 may include N input waveguides 152 and N output waveguides 156’, thus […] to a plurality of output ports of the wavelength router […]is disclosed because Arash teaches that the optical interference unit converts the optical input vector into a second array of optical signals and that an array of
N
output waveguides guides those signals. The N output waveguides correspond to the plurality of output ports, and guiding the second array of optical signals through the output waveguides corresponds to directing the second optical signals to the plurality of output ports)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Wang’s approach of sending a plurality of optical signals using a light source array and routing the optical signals based on wavelengths using a wavelength router to perform computation with Arash’s approach of using a modulator array coupled to a plurality of input ports and a plurality of output ports for receiving at the plurality of input ports and routing to the plurality of output ports, thereby reducing power consumption by leveraging the modulator array architecture optimized for efficient optical signal modulation (Arash, Page 37 – line 12, “In some embodiments, the modulator of the modulator array 144 and/or the OMM unit 150 can be designed to reduce power consumption, so that when the modulator is operated to generate a modulation value representing a more frequently occurring coefficient It consumes less power while operating, and consumes more power when operating the modulator to generate modulation values representing coefficients that occur less frequently. For example, for certain data sets that are known to have certain characteristics, power consumption can be reduced. Figure 42 shows the modulation value probability distribution diagram 4200 (dotted line) superimposed on the modulator power diagram 4202 (solid line) for the specific design of the modulator and/or OMM unit 150 of the modulator array 144”, & Page 38 – line 4, “The power distribution shown in Figure 42 shows that the zero modulation power is used to achieve the zero modulation value, but in other embodiments, there may be residual low but non-zero modulation power at the zero modulation value. For these low coefficient weighted data sets, the power consumption can usually be reduced by using a modulator”, thereby reducing power consumption by leveraging the modulator array optimized for efficient optical signal modulation is disclosed because Arash teaches that the modulator array can be designed such that modulation values corresponding to more frequently occurring coefficients consume less power, resulting in an overall reduction in power consumption for optical signal processing)
Wang combined with Arash does not explicitly disclose […] based on both a wavelength of each intermediate optical signal and an identity of one of the plurality of input ports at which the intermediate optical signal is received […] are routed to the same output port […], […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data […], and […] with an equal sequence number difference (i-j) […].
However, Roberto teaches:
[…] based on both a wavelength of each intermediate optical signal and an identity of one of the plurality of input ports at which the intermediate optical signal is received […] are routed to the same output port […] (Roberto, Page 1 – Section 1, “An arrayed waveguide grating router (AWGR) [7], [8] is an example of devices with such wavelength routing capability. As Fig. 1(a) and (b) illustrate, the well-known wavelength routing property of an AWGR allows any input port to communicate with any output port simultaneously using different wavelengths without contention”, & Page 2 – Fig 1 a and b,
PNG
media_image1.png
412
489
media_image1.png
Greyscale
,
& Page 8 – Section V, “Second, we show a scenario with network contention where C1 and C2 send packets to C3 with random arrival time. To avoid resynchronization in the FPGA GTH receiver, tunable optical delay lines in each transmission path guarantee clock-phase matching of the packets at AWGR output 3”, thus […] based on both a wavelength of each intermediate optical signal and an identity of one of the plurality of input ports at which the intermediate optical signal is received […] are routed to the same output port […] is disclosed because Roberto teaches an AWGR in which the destination output port is determined by the combination of the wavelength used and the input port through which the signal enters, as illustrated by the wavelength-routing property and wavelength-assignment table of Figures 1(a) and 1(b). Roberto further teaches that packets originating from C1 and C2 are directed to the same AWGR output 3)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine Wang and Arash with Roberto by incorporating Roberto’s AWGR wavelength-routing technique into the optical computing system of Wang and Arash. Wang teaches generating wavelength division optical signals that represent products of elements of first and second matrices, while Arash teaches communicating optical signals through arrays of input and output waveguides. Roberto teaches that an AWGR routes signals between input and output ports based on wavelength and permits simultaneous optical communication without contention. Therefore, a POSITA would have been motivated to use Roberto’s AWGR to route Wang’s wavelength encoded product signals from Arash’s input waveguides to selected output waveguides based on the wavelength of each signal and the input port through which the signal is received. This would allow multiple product signals to be routed in parallel, reduce contention, and improve the speed, scalability, and energy efficiency of the optical convolution system (Roberto, Page 1 - Section I, “Compared to electrical interconnects, optical interconnects provide (1) higher transmission bandwidth and lower energy consumption independently of distance, (2) inherent parallelism, (3) low interference and crosstalk, and (4) low parasitic” and “the well-known wavelength routing property of an AWGR allows any input port to communicate with any output port simultaneously using different wavelengths without contention”)
Wang and Arash combined with Roberto does not explicitly […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data […], and […] with an equal sequence number difference (i-j) […].
However, Xu teaches:
[…] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data […] (Xu, Page 3, “The input data vector X is encoded as the intensity of temporal symbols in a serial electrical waveform at a symbol rate 1/τ (baud), where τ is the symbol period. The convolution kernel is similarly represented by a weight vector W of length R that is then encoded in the optical power of the microcomb lines through spectral shaping performed by a Waveshaper. The temporal waveform X is then multi cast onto the kernel wavelength channels via electro-optical modulation, thus generating the replicas weighted by W. Next the optical waveform is transmitted through a dispersive delay with a delay step (between adjacent wavelength channels) equal to the symbol duration of X, effectively achieving time and wavelength interleaving. Finally, the delayed and weighted replicas are summed via high speed photodetection so that each time slot yields a convolution between X and W for a given convolution window, or receptive field. As such, the convolution window effectively slides at the modulation speed matching the baud rate of X. Each output symbol is the result of R multiply-and-accumulate operations”, thus […] and wherein a light intensity at each output port represents one element of a convolution result of the first data and the second data […] is disclosed because Xu teaches that the input vector X, corresponding to the first data, is encoded using optical intensity and that the convolution kernel W, corresponding to the second data, is encoded using the optical power of respective wavelength channels. Xu further teaches generating weighted optical replicas representing products of elements of X and W, aligning and summing those replicas through photodetection, and producing an output symbol for each time slot. Because each output symbol results from the multiply-and-accumulate operations for one convolution window, each output symbol represents one element of the convolution result. In the combined system, applying this intensity-based convolution technique to the plurality of output ports results in the light intensity at each output port representing a respective element of the convolution result)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to further combine Wang, Arash, and Roberto with Xu by incorporating Xu’s intensity based optical convolution and photodetection technique into the combined optical routing system. Wang teaches producing optical signals corresponding to products of elements of two matrices, Arash teaches guiding optical signals through a plurality of output waveguides, and Roberto teaches routing the optical signals to selected output ports. Xu teaches encoding input data and convolution weights using optical intensity and optical power, generating weighted optical replicas, and summing those replicas through high-speed photodetection so that each output symbol represents the multiply-and-accumulate result for a convolution window. Therefore, a POSITA would have been motivated to apply Xu’s photodetection technique to the product signals routed to each output port so that the signals received at that port are summed and represented as a respective convolution-output element. This would provide a direct and high speed technique for obtaining convolution results from the routed optical product signals and would take advantage of the parallel wavelength channels already used by the combined system (Xu, Page 3, “Finally, the delayed and weighted replicas are summed via high speed photodetection so that each time slot yields a convolution between X and W for a given convolution window, or receptive field. As such, the convolution window effectively slides at the modulation speed matching the baud rate of X. Each output symbol is the result of R multiply-and-accumulate operations , with the computing speed given by 2R/τ FLOPS. Since the speed of this process scales with both the baud rate and number of wavelengths, it can be dramatically boosted into the Tera-FLOP regime by using the massively parallel wavelength channels of the microcomb source”)
Wang, Arash, and Roberto combined with Xu does not explicitly […] with an equal sequence number difference (i-j) […].
However, Brea teaches:
[…] with an equal sequence number difference (i-j) […] (Brea, Par. [0192], “Each output vector sample y.sub.p is produced by the time reversal of the matrix sequence {C.sub.p,p=0,1,2, . . . ,M-1} and its circular shifting to the right. The resulting operation shall be called circular matrix convolution. Note that the sum of the indices of each sample product inside the summation modulo M is equal to the index of the output being calculated by the circular matrix convolution”, thus Brea teaches grouping product terms according to a common convolution-output index. Because one of the sequences is time-reversed before the products are formed, the sum-of-indices relationship for the reversed sequence corresponds to a difference
i
j
between the original sequence numbers. Product terms having the same sequence-number difference therefore contribute to the same convolution-output element)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to further combine Wang, Arash, Roberto, and Xu with Brea by using Brea’s known convolution indexing relationship to determine which optical product signals should be routed to the same output port. Wang teaches optical signals representing products of first data and second-data elements, Roberto teaches routing optical signals according to wavelength and input-port identity, and Xu teaches summing product signals to produce individual convolution-output elements. Brea teaches that the product terms contributing to a convolution output are grouped according to a common output index and that one sequence is time reversed before the product terms are formed. Therefore, a POSITA would have been motivated to assign the wavelengths and input ports of the product signals according to Brea’s convolution index relationship so that products having the same sequence number difference (i-j) are directed to the same output port. Xu’s photodetection technique could then sum the grouped product signals at that port to produce the corresponding convolution element. This would implement the known mathematical organization of convolution products directly in the optical routing hardware and allow multiple convolution elements to be calculated efficiently and in parallel (Brea, Paragraph [0192], “Each output vector sample y.sub.p is produced by the time reversal of the matrix sequence {C.sub.p, p=0,1,2, . . . ,M-1} and its circular shifting to the right. The resulting operation shall be called circular matrix convolution. Note that the sum of the indices of each sample product inside the summation modulo M is equal to the index of the output being calculated by the circular matrix convolution”)
Regarding Claim 18, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 17 as cited above and Wang further teaches:
[…] one of the second optical signals comprising at least two intermediate optical signals, wherein each of the at least two intermediate optical signals indicates a product of an element in the first data and an element in the second data (Wang, Page 6-7 – line 44, “It can be understood that the plurality of elements in the first matrix is not loaded on each wavelength of the optical signal according to the time sequence, but one element is loaded on each wavelength, for example: λ1 x1 λ9 loading x9, at this time 9 of the wavelength is loaded with 9 elements. the second modulator is used for loading multiple elements in the second matrix to be processed to each wavelength division optical signal according to the time sequence so as to obtain multiple second modulation optical signals”, & Page 7 – line 4, “for example, the y1, y2 ... y9 are loaded to C1, then obtaining the x1y1, x1y2 ... x1y9, the y1, y2 ... y9 are loaded to C2, then obtaining the x2y1, x2y2 ... x2y9, and so on, y2 ... y9 are loaded to C9, then obtaining x9y1, x9y2 ... x9y9. Similarly, the photoelectric detector only collecting the x1y1, x2y2 ... x9y9 information in the wave-combining optical signal, and converting it into electric signal, finally obtaining the convolution result”, thus one of the second optical signals comprising at least two intermediate optical signals, wherein each of the at least two intermediate optical signals indicates a product of an element in the first data and an element in the second data is disclosed because Wang teaches that elements of the first matrix are loaded onto different wavelengths and that elements of the second matrix are subsequently loaded onto those wavelength-specific optical signals to generate products (e.g., x1y1, x1y2 … x9y9). Wang also discloses that each wavelength carries multiple such product terms, indicating that a given optical signal comprises multiple intermediate optical signals, each representing a product of an element from the first data and an element from the second data)
Wang does not explicitly teach outputting, by a first output port in the output ports […].
However, Arash teaches outputting, by a first output port in the output ports […] (Arash, Page 116 – line 9, “The computing device of claim 17, wherein the replication module includes at least one replication module having an optical splitter, and the optical splitter transmits a predetermined ratio of a power of a light wave at an input port of the replication module To a first output port of the copy module, and send a remaining proportion of the power of the light wave to a second output port of the copy module at the input port of the copy module.”, thus outputting, by a first output port in the output ports is disclosed because Arash teaches an optical splitter within a replication module that receives a light wave at an input port and distributes the light wave to multiple output ports, including a first output port, thereby causing the first output port to output a portion of the light wave)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Wang’s approach of one of the second optical signals comprising at least two intermediate optical signals, wherein each of the intermediate optical signals indicates a product of an element in the first data and an element in the second data, with Arash’s approach of configuring a first output port in a plurality of output ports to output an optical signal, in order to split and distribute the second optical signals comprising the intermediate optical signals to a plurality of output ports (Arash, Page 68- line 20, “the 1×2 optical amplitude modulator includes a ring resonator 2222 configured to divide the optical power of the optical signal from the input port 2221 to two output ports”). Arash further teaches that the modulator array can be designed to reduce power consumption during optical signal modulation (Arash, Page 37 – line 12, “In some embodiments, the modulator of the modulator array 144 and/or the OMM unit 150 can be designed to reduce power consumption, so that when the modulator is operated to generate a modulation value representing a more frequently occurring coefficient It consumes less power while operating, and consumes more power when operating the modulator to generate modulation values representing coefficients that occur less frequently. For example, for certain data sets that are known to have certain characteristics, power consumption can be reduced. Figure 42 shows the modulation value probability distribution diagram 4200 (dotted line) superimposed on the modulator power diagram 4202 (solid line) for the specific design of the modulator and/or OMM unit 150 of the modulator array 144”, & Page 38 – line 4, “The power distribution shown in Figure 42 shows that the zero modulation power is used to achieve the zero modulation value, but in other embodiments, there may be residual low but non-zero modulation power at the zero modulation value. For these low coefficient weighted data sets, the power consumption can usually be reduced by using a modulator”, thereby reducing power consumption by leveraging the modulator array optimized for efficient optical signal modulation, and thereby providing an additional benefit of improved energy efficiency)
Regarding Claim 19, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 17 as cited above and Wang further teaches:
wherein each of the elements corresponds to one of the first optical signals of one wavelength, and wherein each of the first groups comprises first optical signals of different wavelengths (Wang, Page 4 – line 39, “the number of wavelengths of the optical signal output by the multi-wavelength light source 10 according to the number of elements in the first matrix, the number of wavelengths in this example can be 9, then the wavelength corresponding to the elements in the first matrix is respectively represented as λ1 λ9, The first modulator 20 simultaneously loads each element of the first matrix X to 9 wavelengths in the form of light intensity, and the signal loading interval of x1 to x9 is Δ”, & Page 5 – line 1, “In other words, each wavelength is loaded with 9 elements, on each wavelength, the signal loading interval of x1 to x9 is Δ t1, namely, the first modulated light signal can be composed of 9 wavelengths, each wavelength is loaded with 9 elements according to the time sequence”, thus wherein each of the elements corresponds to one of the first optical signals of one wavelength, and wherein each of the first groups comprises first optical signals of different wavelengths is disclosed because Wang teaches that the number of wavelengths corresponds to the number of elements in the first matrix, such that each element is associated with a respective wavelength. Wang further teaches that the optical signal includes multiple wavelengths, and that each wavelength carries signals corresponding to elements, indicating that the first groups include optical signals of different wavelengths)
Regarding Claim 20, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 17 as cited above and Arash further teaches:
detecting, by a detector array in the optical computing apparatus, light intensities of the second optical signals to obtain the computing result (Arash, Page 79 – line 36, “In this embodiment, the optical amplitude modulator array can be replaced by a detector array to convert optical signals into electrical signals, followed by electronic subsystems (such as ASIC, processor, or SoC). Optionally, if optical signal routing is to be used for a summation module configured to detect optical signals, the electronic subsystem may include the use of an array of electrically-modulated optical sources ) Of electro-optical conversion”, & Page 46 – line 3, “The input data set processed by the ANN computing system usually includes data with a resolution greater than 1 bit. For example, ordinary pixels of a gray-scale digital image may have an 8-bit resolution, that is, 256 different levels. One way to represent and process the data in the optical domain is to encode 256 pixels with different intensity levels as 256 different power levels of the optical signal input to the OMM unit 150. Optical signals are analog signals in nature, so they are easily affected by noise and detection errors”, thus detecting, by a detector array in the optical computing apparatus, light intensities of the second optical signals to obtain the computing result is disclosed because Arash teaches a detector array configured to convert optical signals into electrical signals, indicating a plurality of detectors arranged in an array for detecting optical signals. Arash further teaches that optical signals are encoded using different power levels corresponding to different intensity levels, and that such signals are subject to detection, thereby indicating that the detector array detects light intensities of the optical signals to obtain the computing result. Incorporating Arash’s detector array is consistent with Arash’s modulator array optimized for efficient optical signal modulation, which reduces overall power consumption)
Regarding Claim 21, Wang, Arash, Roberto, and Xu combined with Brea teaches all of the limitations of claim 1 as cited above and Arash further teaches:
wherein a quantity of modulators in the modulator array equals a quantity of the plurality of input ports of the wavelength router, wherein each modulator in the modulator array is connected to a corresponding one of the plurality of input ports […](Arash, Page 40, “The optical processor 140 includes a laser unit 142, a modulator array 144, a detection unit 146, and an optical matrix multiplication (OMM) unit 150. The optical processor 140 operates by encoding a digital input vector of length N onto a light input vector of length N and propagating the light input vector through the OMM unit 150”, & Page 40-41, “The OMM unit 150 may include an array of input waveguides 152 to receive light input vectors; an optical interference unit 154 in optical communication with the array of input waveguides 152; and an array of output waveguides 156 in optical communication with the optical interference unit 154. The optical interference unit 154 linearly converts the light input vector into a second optical signal array. The array of output waveguides 156 guides the second array of optical signals output by the optical interference unit 154. At least one input waveguide in the array of input waveguides 152 optically communicates with each output waveguide in the array of output waveguides 156 through the optical interference unit 154. For example, for an optical input vector of length N, the OMM unit 150 may include N input waveguides 152 and N output waveguides 156”, & Page 42, “The modulator array 144 has N modulators, and each modulator receives one of the N light outputs of the laser unit 142. The modulator receives the control signal corresponding to the element of the digital input vector and modulates the intensity of the light”, thus wherein a quantity of modulators in the modulator array equals a quantity of the plurality of input ports of the wavelength router, wherein each modulator in the modulator array is connected to a corresponding one of the plurality of input ports […] is disclosed because Arash teaches a modulator array having N modulators that encodes a digital input vector of length N as a light input vector of length N. Arash further teaches that the OMM unit receives the light input vector through N input waveguides, which correspond to the plurality of input ports. Because the modulator array includes N modulators and the OMM unit includes N input waveguides, the quantity of modulators equals the quantity of input ports. Each modulator produces a respective modulated optical signal representing a corresponding element of the digital input vector, and the resulting N optical signals are provided to the respective N input waveguides, corresponding to each modulator being connected to a corresponding input port)
Brea further teaches:
[…] wherein a quantity of the plurality of output ports of the wavelength router is m + n – 1, wherein m is a quantity of elements in the first data, and wherein n is a quantity of elements in the second data (Brea, Par. [0205], “Linear Matrix Convolution and Correlation: Observe that the circular matrix convolution of M point matrix and vector sequences yields another M-point vector sequence. On the other hand, a linear matrix convolution of an M.sub.1-point matrix sequence with an M.sub.2-point vector sequence would produce an (M.sub.1+M.sub.2-1)-point vector sequence”, […] wherein a quantity of the plurality of output ports of the wavelength router is m + n – 1, wherein m is a quantity of elements in the first data, and wherein n is a quantity of elements in the second data is disclosed because Brea teaches that a linear convolution between an M1-point sequence and an M2-point sequence produces an M1 + M2 – 1-point output sequence. The M1-point sequence corresponds to the first data having m elements, the M2-point sequence corresponds to the second data having n elements, and the resulting M1 + M2 – 1 output points correspond to the m + n – 1 elements of the convolution result. In the combined optical system, providing a respective wavelength-router output port for each convolution-result element results in m + n – 1 output ports)
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office
action. THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MAHLIET ADMASU whose telephone number is (571)272-0034. The examiner can normally be reached Mon-Fri, 8am-5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Alexey Shmatov can be reached at (571)270-3428. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/M.T.A./
Examiner, Art Unit 2123
/ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123
Prosecution Insights