DETAILED ACTION
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 .
This action is in response to the application and claims filed 8/21/2023. Claims 1-20 are pending and have been examined. Claims 1-17 are rejected and claims 18-20 are objected to.
Priority
Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, 365(c), or 386(c) is acknowledged. The present application claims foreign priority to Korean patent application No. KR10-2023-0014612filed on 2/03/2023.
The examiner acknowledges that a certified copy of Korean patent application No. KR10-2023-0014612 has been retrieved (on 9/26/2023, in Korean).
The examiner notes that a translation of Korean patent application No. KR10-2022-0045188 does not appear to have been furnished to-date.
Information Disclosure Statement
Acknowledgment is made of the information disclosure statements filed 8/21/2023 and 11/13/2025, which comply with 37 CFR 1.97. As such, the information disclosure statements have been placed in the application file and the information referred to therein has been considered by the examiner.
Drawings
The drawings are objected to as failing to comply with 37 CFR 1.84(p)(5) because they include the following reference characters not mentioned in the description:
Reference characters ss12, ss13, ss32 and ss33 shown in Figure 3 are not found in the detailed description (see, paragraphs 66-87 describing FIG. 3); and
Reference characters ON_4, os2 and os4 shown in Figures 5A, 5B and 5C are not found in the detailed description (see, paragraphs 98-115 describing FIGs. 5A-5C).
The drawings are further objected to as failing to comply with 37 CFR 1.84(p)(3) because Figures 1-7 include letters which do not measure at least .32 cm. (1/8 inch) in height (i.e., most of the lowercase characters in FIGs. 1-7 and most of the lowercase and subscript characters in FIG. 4).
Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. The figure or figure number of an amended drawing should not be labeled as “amended.” If a drawing figure is to be canceled, the appropriate figure must be removed from the replacement sheet, and where necessary, the remaining figures must be renumbered and appropriate changes made to the brief description of the several views of the drawings for consistency. Additional replacement sheets may be necessary to show the renumbering of the remaining figures. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
Specification
The disclosure is objected to because of the following informalities:
Reference characters ss12, ss13, ss32 and ss33 shown in Figure 3 are not found in the detailed description (see, paragraphs 66-87 describing FIG. 3).
Reference characters ON_4, os2 and os4 shown in Figures 5A, 5B and 5C are not found in the detailed description (see, paragraphs 98-115 describing FIGs. 5A-5C).
Appropriate correction is required.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f):
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f). The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f), is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f). The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f), is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f), except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f), except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f), because the claim limitations use a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitations are:
a sampler configured to generate session data and
a spike signal generator configured to generate first to m-th spike signals in claim 1.
Regarding claim 1 and the above-noted three-prong test, the recited sampler is a generic placeholder, configured to generate session data is functional language, and there is no recitation in claim 1 of sufficient structure to perform the generating. Also in claim 1, the recited spike signal generator is a generic placeholder, configured to generate first to m-th spike signals is functional language, and there is no recitation in claim 1 of sufficient structure to perform the generating.
Because these claim limitations are being interpreted under 35 U.S.C. 112(f), they are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
A review of the specification shows that the corresponding structure is not described in the specification for the 35 U.S.C. 112(f) limitations:
Regarding the above-noted sampler and spike signal generator recited in claim 1, although a sampler and spike signal generator are depicted in the black box block diagram of FIG. 1 and generally mentioned in paragraphs 5, 11, 13-14, 17 and 22 (which merely repeat the claim language) and 38, 41-43 and 46 (which generally describe the aforementioned figure), the corresponding structure of the claimed sampler and spike signal generator capable of performing the claimed functions is not described in applicant’s specification.
The drawings merely show black-boxes designed to perform the entire claimed function (see, e.g., “Sampler” 20 and “Spike Signal Generator” 30 shown in FIG. 1).
As such, the specification either fails to describe the claimed sampler and spike signal generator as noted above, or, at best, describes the claimed sampler and spike signal generator by their respective functions without disclosing any specific structure performing the claimed functions.
Accordingly, for these claim limitations, the written description fails to disclose both an algorithm(s) and special-purpose computer hardware to perform the algorithm(s). For more information, see MPEP § 2181.
If applicant wishes to provide further explanation or dispute the examiner's interpretation of the corresponding structure, applicant must identify the corresponding structure with reference to the specification by page and line number, and to the drawing, if any, by reference characters in response to this Office action.
If applicant does not intend to have these limitations interpreted under 35 U.S.C. 112(f), applicant may: (1) amend the claim limitations to avoid them being interpreted under 35 U.S.C. 112(f) (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitations recite sufficient structure to perform the claimed function so as to avoid them being interpreted under 35 U.S.C. 112(f).
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1-12 are rejected under 35 U.S.C. 112(a) as failing to comply with the written description requirement.
Independent claim 1 and dependent claims 2-12 each contain subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, at the time the application was filed, had possession of the claimed invention.
In particular, and as previously noted, the claim limitations a sampler configured to generate session data and a spike signal generator configured to generate first to m-th spike signals in claim 1 invoke 35 U.S.C. 112(f).
However, as noted above, the written description of the current application fails to disclose the corresponding structure, material, or acts for performing each of the above-identified claimed functions and to clearly link the structure, material, or acts to the function. In particular, for each of the claimed functions, the written description fails to disclose both an algorithm(s) and special-purpose computer hardware to perform the algorithm. For more information, see MPEP § 2181.
Accordingly, claim 1 is rejected under 35 U.S.C. 112(a) as failing to comply with the written description requirement.
Also, claims 2-12, which each depend directly or indirectly from claim 1, are rejected under 35 U.S.C. 112(a) as failing to comply with the written description requirement under the same rationale as claim 1.
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 1-12 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
As discussed above, the claim limitations a sampler configured to generate session data and a spike signal generator configured to generate first to m-th spike signals in claim 1 invoke 35 U.S.C. 112(f).
However, as also discussed above with regard to the rejections of independent claim 1 under 35 U.S.C. 112(a), the written description fails to disclose the corresponding structure, material, or acts for performing the entire claimed function and to clearly link the structure, material, or acts to the function. In particular, the specification fails to clearly link the structure, material, or acts to the function for the limitations a sampler configured to generate session data and a spike signal generator configured to generate first to m-th spike signals in claim 1.
As further noted above, there is insufficient disclosure in the specification of algorithms and specific computer hardware for implementing the above-noted, claimed sampler and spike signal generator. As such, the above-noted limitations recited in claim 1 are indefinite. Therefore, claim 1 is indefinite and are rejected under 35 U.S.C. 112(b). For the purposes of determining patent eligibility and comparison with the prior art, the examiner is interpreting the above-listed sampler and spike signal generator as any combination of software (i.e., a set of instructions, code, one or more functions or software modules) and/or hardware (i.e., circuitry and/or hardware logic components/modules) capable of performing the claimed functions.
Also, claims 2-12, which each depend directly or indirectly from independent claim 1, are rejected under 35 U.S.C. 112(b) as being indefinite under the same rationale as claim 1.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-6 and 13-17 are rejected under 35 U.S.C. 103 as being unpatentable over Verzi et al (U.S. Patent Application Pub. No. 2022/0406408 A1, hereinafter “Verzi”) in view of Moraitis et al. (U.S. Patent Application Pub. No. 2020/0184325 A1, hereinafter “Moraitis”).
Regarding claim 1, Verzi discloses the invention as claimed including an anomaly data detection device (see, e.g., paragraph 8, “a system for anomaly detection for streaming data. The system comprises a storage device”) comprising:
a sampler1 configured to generate session data including first to m-th sample data based on input data input during a first time interval (see, e.g., paragraphs 43, “Under … spike-timing-dependent plasticity (STDP) the weight connecting pre- and post-synaptic units is adjusted according to their relative spike times within a specified time interval.”, 52-53, “spiking neurons 410, 414, and 418 in spiking neural architecture 400 might be initialized using … each input value in relation to all other possible input values {x1 , x2 , ... , xp} … spiking neurons thus provide inhibitory signals such that the first one … of its originally associated input signal xi, corresponding to the sample median of the original array of input values”, “In spiking neural architecture 400, each input x, is connected to each spiking neuron ni, … if multiple input values correspond to the same value as the sample median, then all of their associated spiking neurons will spike simultaneously.” and 63, “a data stream (e.g., sequential, time-series) … can comprise a specified number of input values before and after a reference input or input values within a specified time frame” [i.e., sampler to generate session/data stream including x1/1st to xm/mth sample input values from array of input values during a 1st time interval/frame]);
a spike signal generator2 configured to generate first to m-th spike signals respectively corresponding to the first to m-th sample data based on the session data (see, e.g., paragraphs 42, “communication in SNNs is done by broadcasting trains of action potentials, known as spike trains. … a spike is generated when the sum of changes in a neuron's membrane potential resulting from pre-synaptic stimulation crosses a threshold. This principle is simulated in artificial SNNs in the form of a signal accumulator that fires when a certain type of input surpasses a threshold.”, 45, “Accumulation of a series of input spikes”, 48, “energy provided by inputs until a threshold is reached and the neuron fires as a spike that provides input to other neurons via synapse connections. By emitting this spike, the neuron is returned to a low energy state and continues to integrate input current until its next firing.” and 56, “when neuron j fires … xi; are real valued inputs; zq are phase-coded spike inputs with a phase-code value of pq (from other phase-coded spiking neurons)” [i.e., generate/emit 1st to mth spikes/spike signals corresponding to the sampled input values]);
a spike neural network configured to detect whether an output spike fires in at least one output neuron from among a plurality of output neurons based on the first to m-th spike signals and synaptic weights of each of the plurality of output neurons (see, e.g., paragraphs 41, “a Spiking Neural Network (SNN)”, 43, “Information in SNNs is conveyed by spike timing, including latencies and spike rates … Under … spike-timing-dependent plasticity (STDP) the weight connecting pre- and post-synaptic units is adjusted according to their relative spike times within a specified time interval.”, 47, “Firing: Emission of an output spike when the accumulated signal reaches a certain level” and 53-54, “In spiking neural architecture 400, each input xi is connected to each spiking neuron ni, where the weights are set to wij=sign(xi-xj)/xj”, “a multi-layer, spiking, adaptive median-filtering (AMF) network” [i.e., a spiking neural network/SNN detects whether an output spike fires in spiking neuron ni based on the accumulated 1st to mth spike signals and synaptic weights wij of the neurons]); and
a detection circuit configured to generate a detection signal based on the number of output neurons firing the output spike (see, e.g., paragraphs 32, “the hardware may take the form of a circuit system, an integrated circuit, an application-specific integrated circuit (ASIC)”, 63, “In the case of a data stream (e.g., sequential, time-series), a neighborhood can comprise a specified number of input values before and after a reference input or input values within a specified time frame before and after the reference input.” and 72-73, “AMF [spiking, adaptive median-filtering] layer 550 compares the MAD [median absolute difference] values of a data cell, represented by a number of spiking neurons rm 552 for each size neighborhood (i.e., rij1, rij2, rij3, rij4) to predefined threshold values … and an anomaly is detected, causing the representative neuron rm for that neighborhood to spike. If an anomaly is detected in any size neighborhood, the median value of the data cell for the smallest neighborhood, represented by phase-coded spike pij1, is output by the AMF layer 550”, “If none of the neurons 552 spike, no anomaly is detected” [i.e., a detection circuit generates an anomaly detection output signal based on the number of neurons firing/spiking]).
Although Verzi substantially discloses the claimed invention and Verzi discloses “Each visible node in layer 310 takes a low-level feature from an item in the dataset and passes it to the hidden nodes in the next layer” and “Information in SNNs is conveyed by spike timing, including latencies and spike rates.” (see, paragraphs 39 and 43), Verzi is not relied on for explicitly disclosing wherein each of the first to m-th spike signals is generated by converting feature information of the corresponding first to m-th sample data into a spike rate code.
However, in the same field, analogous art Moraitis teaches wherein each of the first to m-th spike signals is generated by converting feature information of the corresponding first to m-th sample data into a spike rate code (see, e.g., paragraphs 26, “the neuron at the given time bin generates a spike … where … Rk is the determined spike rate of the neuron k. Thus, at each point in time, the probability of spike generation … depends on a spike rate that is specific to each neuron” [i.e., generate spikes/spike signals based on a spike rate for each 1st-mth neurons], 34, “features of the long timescale may be represented by … the spike rate”, 56, “the rate coding operates at a slower timescale than time-based coding, and it does not determine a unique choice of the individual spikes that make up the rate-coded pattern e.g. different sets of individual spikes may be used for providing the rate pattern of a same variable value. … the same rate value may be obtained by different sets of individual spikes. The rate defined by a selected set of individual spikes that would also comprise a spike having a timing that encodes the other variable value may be used for spike firing.” and 105, “spike-rate component 515 computes the rates of the presynaptic input spikes.” [i.e., generate spike signals/spikes for each 1st-mth neuron by converting features/variable values of the corresponding selected/sampled data into determined/computed spike rate codes/rates for each neuron]).
Verzi and Moraitis are analogous art because they are both directed to techniques and systems for using and implementing spiking neural networks/SNNs (see, e.g., Verzi, Abstract and paragraphs 7-9, and Moraitis, Abstract and paragraphs 6-7, 23 and 25).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the disclosed system of Verzi to incorporate the teachings of Moraitis to provide “a method of generating spikes by a neuron of a spiking neural network. … wherein the spike generation encodes at each time instant at least two variable values [i.e., features] at the neuron. Synaptic weights may be optimized for a spike train generated by a given presynaptic neuron of a spiking neural network, wherein the spike train being indicative of features of at least one timescale” and to “enable [the spiking neural network] to process data features of short timescales separately from but simultaneously with those features of long timescales. Features of the short timescale may be represented by individual spikes, and features of the long timescale may be represented by multi-spike sequences of longer patterns, such as in the spike rate.” (see, e.g., Moraitis, Abstract and paragraph 34). One of ordinary skill in the art would have been motived to combine the system of Verzi with the spiking neural network and spike rate of Moraitis because the spike rate enables the Moraitis spiking neural network “to process data features of short timescales separately from but simultaneously with those features of long timescales. This is particularly advantageous as temporal or spatiotemporal data often include features in multiple timescales”, as suggested by Moraitis. (see, e.g., Moraitis, paragraph 34).
Regarding claim 2, as discussed above, Verzi in view of Moraitis teaches the device of claim 1.
Although Verzi substantially discloses the claimed invention and Verzi discloses that “In the case of a data stream (e.g., sequential, time-series), a neighborhood can comprise a specified number of input values before and after a reference input or input values within a specified time frame before and after the reference input. … This hierarchical arrangement of anomaly detection is even more important when handling streaming data (e.g., … finding multiple of these in larger neighborhood sizes” (see, paragraph 63), Verzi is not relied on for explicitly disclosing wherein the feature information includes a size value and a change value of each of the first to m-th sample data.
However, in the same field, analogous art Moraitis teaches wherein the feature information includes a size value and a change value of each of the first to m-th sample data (see, e.g., paragraphs 34, “process data features of short timescales separately from but simultaneously with those features of long timescales … features of the long timescale may be represented by multi-spike sequences of longer patterns”, 75, “a change dx of input signal amplitude that is associated with a shortest relevant timescale dt (referred to as short timescale) in each sequence is encoded as a unitary event or spike. … features of the short timescale dt are represented by the timing of spikes in the input sequences. In each of the input sequences, patterns of a timescale longer than dt (long timescale) can be described by the number of spikes within a longer temporal window of size Dt” and 96, “brightness may be represented in a short timescale dt of the spiking data, e.g. by the timing of individual spikes. A long timescale can, for example, be described by the number of spikes within a longer temporal window of size ∆t, longer than dt, or, equivalently, the rate measured in window ∆t, which here represents the pixel's color value. [i.e., data features/feature information of sample input data includes a window size and change values dt/∆t/delta of the data]).
The motivation to combine Verzi and Moraitis is the same as discussed above with respect to claim 1.
With respect to independent claim 13, claim 13 is substantially similar to claim 1 and therefore is rejected on the same ground as claim 1, discussed above. In particular, claim 13 is a method claim with steps that correspond to the device operations of claim 1.
In addition, Verzi further discloses a method of operating an anomaly data detection device including a sampler, a spike signal generator, a spike neural network, and a detection circuit (see, e.g., paragraphs 7, “a computer-implemented method of anomaly detection for streaming data. The method comprises implementing a spiking neural network that performs the steps of: receiving inputs of streaming data, wherein each input is contained within a number of neighborhoods, wherein the neighborhoods comprise a number of increasingly larger symmetrical numbers of inputs preceding and following the input; converting the inputs into phase-coded spikes; calculating, from the phase-coded spikes, a median value of each input … calculating an absolute difference of each input from its median value for each size neighborhood containing the input; calculating, from the absolute differences, a median absolute difference (MAD) value of each input for each size neighborhood containing the input; determining for each input whether the MAD value for any size neighborhood containing the input exceeds a respective threshold, wherein; responsive to a determination that the MAD value of one or more neighborhoods exceeds its threshold, outputting an anomaly indication for the input” and 32, “the hardware may take the form of a circuit system, an integrated circuit, an application-specific integrated circuit (ASIC)” [i.e., a method for operating an anomaly detection device including the sampler, spike signal generator, spike neural network, and detection circuit as discussed above with reference to the sampler, spike signal generator, spike neural network, and detection circuit recited in claim 1]).
Regarding claims 3 and 14, as discussed above, Verzi in view of Moraitis teaches the device of claim 1 and the method of claim 13.
Although Verzi substantially discloses the claimed invention, Verzi is not relied on for explicitly disclosing wherein the spike neural network determines the at least one output neuron to detect whether the output spike fires based on a first correlation between a random distribution of the first to m-th spike signals and a random distribution of the synaptic weights connected to each of the plurality of output neurons.
However, in the same field, analogous art Moraitis teaches wherein the spike neural network determines the at least one output neuron to detect whether the output spike fires based on a first correlation between a random distribution of the first to m-th spike signals and a random distribution of the synaptic weights connected to each of the plurality of output neurons (see, e.g., paragraphs 26, “at each point in time, the probability of spike generation depends on the neuron's synchrony and is conditional on the reference Poisson process which depends on a spike rate that is specific to each neuron and a source of random numbers that is common to all neurons. The common random number allows introduction of synchrony between the neurons”, 48, “the randomness required to implement … the multi-dimensional neuron, can be applied for example to randomly tune the firing threshold around a mean value, or to tune one or both of the input signals.”, 61, “generating a random number using the source 109 from a continuous uniform distribution … and comparing this to … the instantaneous rate (of the range Rmin-Rmax) associated with a given variable value received at the neuron k. In other terms, the probability of spike release may be determined at each time bin dt for each neuron” and 100, “All weights are initialized randomly and uniformly.” [i.e., the spiking neural network/SNN determines the output neuron to detect whether the output spike fires based on a first correlation between a random distribution of the 1st to m-th spikes/spike signals and a random distribution of weights connected to each output neuron]).
Verzi and Moraitis are analogous art because they are both directed to techniques and systems for using and implementing spiking neural networks/SNNs (see, e.g., Verzi, Abstract and paragraphs 7-9, and Moraitis, Abstract and paragraphs 6-7, 23 and 25).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the disclosed system of Verzi to incorporate the teachings of Moraitis to provide “a method of generating spikes by a neuron of a spiking neural network. … wherein the spike generation encodes at each time instant at least two variable values [i.e., features] at the neuron. Synaptic weights may be optimized for a spike train generated by a given presynaptic neuron of a spiking neural network, wherein the spike train being indicative of features of at least one timescale” and to “enable [the spiking neural network] to process data features of short timescales separately from but simultaneously with those features of long timescales. Features of the short timescale may be represented by individual spikes, and features of the long timescale may be represented by multi-spike sequences of longer patterns, such as in the spike rate.” (see, e.g., Moraitis, Abstract and paragraph 34). One of ordinary skill in the art would have been motived to combine the system of Verzi with the spiking neural network and spike rate of Moraitis because the spike rate enables the Moraitis spiking neural network “to process data features of short timescales separately from but simultaneously with those features of long timescales. This is particularly advantageous as temporal or spatiotemporal data often include features in multiple timescales”, as suggested by Moraitis. (see, e.g., Moraitis, paragraph 34).
Regarding claims 4 and 15, as discussed above, Verzi in view of Moraitis teaches the device of claim 1 and the method of claim 13.
Although Verzi substantially discloses the claimed invention, Verzi is not relied on for explicitly disclosing wherein the spike neural network performs a first operation on the synaptic weights connected to the at least one output neuron and the first to m-th spike signals,
wherein the spike neural network performs a second operation on a result of the first operation and the at least one output neuron, and
wherein the spike neural network detects whether the output spike is fired in the at least one output neuron when a result of the second operation exceeds a threshold value.
However, in the same field, analogous art Moraitis teaches wherein the spike neural network performs a first operation on the synaptic weights connected to the at least one output neuron and the first to m-th spike signals (see, e.g., paragraphs 96-97, “2D neurons may be used at the input layer of a SNN network … The 2D neuron may be configured to generate spikes … to encode the two values in the generated spikes”, “The SNN may be formed of 100 2D neurons (presynaptic neurons) and the two output neurons. One output neuron may receive weights from short-term components of the 100 neurons and the other output neuron may receive weights from long-term components of 100 neurons”, 100-101, “All weights are initialized randomly and uniformly … the initial weight distribution of the input synapses to the long-timescale and the short-timescale neurons … receiving input … of spiking data”, “The SNN is trained using 20 examples of each class of the classes” and 104, “network (SNN) 500 learning from spike inputs” [i.e., the spiking neural network/SNN performs a 1st operation on the weights connected to the output neuron and the spike signals/spike inputs]),
wherein the spike neural network performs a second operation on a result of the first operation and the at least one output neuron (see, e.g., paragraphs 102, “the trained network's classification performance is tested by presenting it 5 test images of each class and counting the output neurons' spikes.” and 105-106, “The long-term component outputs a set of aimed synaptic weights 516 during operation of the network. The spike-rate component 515 computes the rates of the presynaptic input spikes. The long-term component sets the aimed weights 516 according to the input spikes”, “The neuron's STDP synapses compare the inputs directly with the neuron's output, so a neuron's learning is sensitive specifically to the uncentered covariance of the inputs.” [i.e., the spiking neural network/SNN performs a 2nd operation on a result of the 1st operation on the weights and the output neuron]), and
wherein the spike neural network detects whether the output spike is fired in the at least one output neuron when a result of the second operation exceeds a threshold value (see, e.g., paragraphs 2, “neurons are provided with a firing threshold that must be exceeded by a membrane potential of the neurons in order to generate a spike. This thresholding is a component of spiking neural network (SNNs).” and 48, “The memristive neuron can comprise a memristor that resets its state when it reaches a threshold value, each reset representing a spike. The neuron may receive two input signals, one signal modulating the rate of the produced spikes as measured in a time window, and another signal modulating the exact timing of the produced spikes. … spikes may be produced by a capacitor that discharges when it reaches a threshold. … the multi-dimensional neuron, can be applied for example to randomly tune the firing threshold around a mean value, or to tune one or both of the input signals.” [i.e., the spiking neural network/SNN detects an output spike firing in the neuron when the 2nd operations exceeds the firing threshold]).
Verzi and Moraitis are analogous art because they are both directed to techniques and systems for using and implementing spiking neural networks/SNNs (see, e.g., Verzi, Abstract and paragraphs 7-9, and Moraitis, Abstract and paragraphs 6-7, 23 and 25).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the disclosed system of Verzi to incorporate the teachings of Moraitis to provide “a method of generating spikes by a neuron of a spiking neural network. … wherein the spike generation encodes at each time instant at least two variable values [i.e., features] at the neuron. Synaptic weights may be optimized for a spike train generated by a given presynaptic neuron of a spiking neural network, wherein the spike train being indicative of features of at least one timescale” and to “enable [the spiking neural network] to process data features of short timescales separately from but simultaneously with those features of long timescales. Features of the short timescale may be represented by individual spikes, and features of the long timescale may be represented by multi-spike sequences of longer patterns, such as in the spike rate.” (see, e.g., Moraitis, Abstract and paragraph 34). One of ordinary skill in the art would have been motived to combine the system of Verzi with the spiking neural network and spike rate of Moraitis because the spike rate enables the Moraitis spiking neural network “to process data features of short timescales separately from but simultaneously with those features of long timescales. This is particularly advantageous as temporal or spatiotemporal data often include features in multiple timescales”, as suggested by Moraitis. (see, e.g., Moraitis, paragraph 34).
Regarding claims 5 and 16, as discussed above, Verzi in view of Moraitis teaches the device of claim 4 and the method of claim 15.
Although Verzi substantially discloses the claimed invention and Verzi discloses “The strength of the synapses (weights) can be changed as a result of learning” and “If a pre-synaptic unit fires before the post-synaptic unit within the specified time interval, the weight connecting them is increased” (see, paragraphs 42-43), Verzi is not relied on for explicitly disclosing wherein the spike neural network increases the synaptic weights connected to the at least one output neuron when the output spike is fired in the at least one output neuron.
However, in the same field, analogous art Moraitis teaches wherein the spike neural network increases the synaptic weights connected to the at least one output neuron when the output spike is fired in the at least one output neuron (see, e.g., paragraphs 32, “tuning the synaptic weight W(t) … The efficacy is a weight. It is named "efficacy" because the increase in the value of the efficacy may induce an increase in the firing probability and thus a firing efficiency of a postsynaptic neuron that receives inputs weighted”, 110, “The STDP rule component 520 may modify the weight component 511 according to the timing of spikes from the pre- and the post-synaptic neurons. … G(t)-W(t). F(t) can be implemented, in an example, in a spike-based way as a function that increases its value by a fixed amount upon the arrival of a presynaptic spike … W(t) denotes the stored synaptic weight” [i.e., the spiking neural network/SNN increases the synaptic weights W(t) value when an output spike is fired]).
The motivation to combine Verzi and Moraitis is the same as discussed above with respect to claims 4 and 15.
Regarding claims 6 and 17, as discussed above, Verzi in view of Moraitis teaches the device of claim 4 and the method of claim 15.
Although Verzi substantially discloses the claimed invention, and Verzi discloses “If a pre-synaptic unit fires before the post-synaptic unit within the specified time interval, the weight connecting them is increased … If it fires after the post-synaptic unit within the time interval, the weight is decreased” (see, paragraph 43), Verzi is not relied on for explicitly disclosing wherein the output neurons include the at least one output neuron and remaining output neurons, and
wherein, when the output spike is fired in the at least one output neuron, the spike neural network maintains synaptic weights connected to each of the remaining output neurons.
However, in the same field, analogous art Moraitis teaches wherein the output neurons include the at least one output neuron and remaining output neurons (see, e.g., paragraph 97, “The SNN may be formed of 100 2D neurons (presynaptic neurons) and the two output neurons. One output neuron may receive weights from short-term components of the 100 neurons and the other output neuron may receive weights from long-term components of 100 neurons”), and
wherein, when the output spike is fired in the at least one output neuron, the spike neural network maintains synaptic weights connected to each of the remaining output neurons (see, e.g., paragraphs 97, “One output neuron may receive weights from short-term components of the 100 neurons and the other output neuron may receive weights from long-term components of 100 neurons” and 105-106, “network 500 comprises an apparatus 510 having a long-term synaptic weight component. The long-term component may comprise a spike-rate component 515 and at least one synapse 511. The long-term component outputs a set of aimed synaptic weights 516 during operation of the network. The spike-rate component 515 computes the rates of the presynaptic input spikes. The long-term component sets the aimed weights 516 according to the input spikes”, “in a spiking neural network, the neurons need to learn to associate the strongly covarying inputs. … In STDP, a neuron approximates this by potentiating, i.e., strengthening, and depressing, i.e., weakening, the synapses whose input shows high and low covariance respectively with the neuron's own output, i.e. the weighted sum of the input streams.” [i.e., when the output neuron fires/spikes, the spiking neural network/SNN maintains/sets a set of synaptic weights connected to the other output neuron]).
The motivation to combine Verzi and Moraitis is the same as discussed above with respect to claims 4 and 15.
Allowable Subject Matter
Upon overcoming of all the rejections as discussed above in items 10-18, claims 7-12 and 18-20 are objected to as being dependent upon a rejected base claim (i.e., independent claims 1 and 13, respectively), but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims (i.e., claim 7 in the case of claims 8-9, claim 10 in the case of claims 11-12 and claim 18 in the case of claims 19-20).
For example, with regard to dependent claims 7-12 and 18-20, the prior art of record does not anticipate, nor do they render obvious in any reasonable combination to one of ordinary skill in the art at the time of Applicants' invention, the combination of recited limitations of claims 7, 10 and 18, and their respective base claims, independent claims 1 and 13, respectively.
As discussed above, Verzi in view of Moraitis teaches the device of claim 1 and the method of claim 13.
Verzi also discloses that “Under … spike-timing-dependent plasticity (STDP) the weight connecting pre- and post-synaptic units is adjusted according to their relative spike times within a specified time interval. If a pre-synaptic unit fires before the post-synaptic unit within the specified time interval, the weight connecting them is increased (long-term potentiation (LTP)). If it fires after the post-synaptic unit within the time interval, the weight is decreased (long-term depression (LTD)).” (see, e.g., paragraph 43).
However, the prior art of record does not anticipate or render obvious the limitations “wherein the sampler generates first session data based on first input data input during the first time interval, and generates second session data based on second input data input during the first time interval, and
wherein the spike neural network detects whether the output spike fires in at least one first output neuron with respect to the first session data or the second session data based on a second correlation between the first session data and the second session data”, as recited, using respective similar language, in claims 7 and 18, in combination with limitations of base claims 1 and 13.
The prior art of record also does not anticipate or render obvious the limitations “wherein the sampler generates fourth session data based on fourth input data input during a second time interval after the first time interval,
wherein the spike neural network detects whether the output spike is fired in the at least one output neuron with respect to the fourth session data, and
wherein the detection circuit generates a first detection signal based on an absence of an output neuron firing the output spike above a first value”, as recited in claim 10, in combination with limitations of base claim 1.
Conclusion
The prior art made of record, listed on form PTO-892, and not relied upon, is considered pertinent to applicant's disclosure.
The references listed on form PTO-892 are all generally related to techniques, methods and systems for implementing and using spiking neural networks/SNNs for machine learning tasks such as anomaly detection.
For example, non-patent literature Stratton et al. ("A spiking neural network based auto-encoder for anomaly detection in streaming data." 2020 IEEE Symposium Series on Computational Intelligence (SSCI). IEEE, 2020), hereinafter “Stratton”) discloses “Automating the detection of anomalies is necessary to deal with large volumes of data and to satisfy real time processing constraints. … Spiking Neural Networks (SNNs), an emerging ML technique, have the potential to do AD [anomaly detection] well … we investigate SNNs doing anomaly detection on streams of text. We show that SNNs are well suited for detecting anomalous character sequences, that they can learn rapidly, and that there are many optimizations to the SNN architecture and training that can improve AD performance.” (see, Abstract).
The examiner requests, in response to this office action, support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line no(s) in the specification and/or drawing figure(s). This will assist the examiner in prosecuting the application.
When responding to this office action, Applicant is advised to clearly point out the patentable novelty which he or she thinks the claims present, in view of the state of the art disclosed by the reference cited or the objections made. He or she must also show how the amendments avoid such references or objections See 37 CFR 1.111 (c).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RANDY K BALDWIN whose telephone number is (571)270-5222. The examiner can normally be reached on Mon - Fri 9:00-6:00.
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, Kamran Afshar can be reached at (571) 272-7796. 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.
/RANDALL K. BALDWIN/Primary Examiner, Art Unit 2125
1 As indicated above in the section 112(b) rejection of this claim, the sampler has been interpreted to be any combination of software (i.e., a set of instructions, code, one or more functions or software modules) and/or hardware (i.e., circuitry and/or hardware logic components/modules) capable of performing the claimed functions.
2 As indicated above in the section 112(b) rejection of this claim, the spike signal generator has been interpreted to be any combination of software (i.e., a set of instructions, code, one or more functions or software modules) and/or hardware (i.e., circuitry and/or hardware logic components/modules) capable of performing the claimed functions
Prosecution Insights