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 Amendment/Arguments
Claims 21-27, 34-46 are pending per 04/30/2026.
A new ground of rejection has been established to addressed the amended language “the encoding neural network comprises a first neural network and a second neural network, and the first neural network reuses at least a part of neuron structures of a neural network that is the same as a decoding neural network that corresponds to the first data stream”.
Thus, arguments pertaining the encoder of O’Shea not having a first and a second neural network is moot in view of the new ground of rejection.
In response to Applicant’s arguments pertaining the limitation that recites reusing neuron structure of the decoder:
Applicant argues O’Shea describes one or more same layer(s) serving both encoder and decoder, rather than encoder having being internally separate neural network and reuses neuron structure of the decoder.
The examiner respectfully disagrees, by at least in view of the ambiguity and lack of technical clarity of the term “reuse”. Applicant has imposed a much narrower interpretation for the limitation than the broad claim language permits. Note that the claim offers no technical definition or otherwise distinct meaning to the term “reuse” to limit claim interpretation in certain way. In plain meaning in light of the context of the claim, the act of reusing under BRI can be any of borrowing, reclaiming, sharing a same structure/entity/object. Given that in ¶0073 of O’Shea, the encoder 202 uses a one or more layers that make up the decoder 204, it can be reasonably constitute the act of reusing, i.e. borrowing or sharing or otherwise using again the same structure.
Furthermore, nothing in the claims explicitly excludes the second neural network of the encoder also reusing at least parts of the decoder. Nothing in O’Shea excludes the encoder/decoders 202 and 204 from having sub networks given they can have multiple layers per ¶0073. As neuron structures of a neural network are defined by parameters, weight values and layers, ¶0073 discloses parameters and weight values are replicated throughout the fully connected neural network system, it can reasonably be said that the encoder network at least partly reuse neuron structure of the decoder. Note that Applicant’s published Specification gives some examples of reuse case wherein ¶0013-0014 disclose: “ the first neural network uses all neuron structures of the decoding neural network” or “the first neural network may reuse a part of the decoding neural network (and/or a parameter of a part of the decoding neural network)”, which fits with the disclosure of ¶0073 of O’Shea.
In view of the foregoing, the arguments are not persuasive.
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
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.
Claim(s) 21-27, 34-46 is/are rejected under 35 U.S.C. 103 as being unpatentable over O’Shea et al. (US 2018/036192) and in view of Van den Oord et al. (US 2018/0025257).
As to claim 21, 41:
O’Shea discloses:
A method, and a non-transitory computer readable medium (Abstract, ¶0014, method and CRM) storing instructions when executed by a processor to cause a transmit end to perform said method, the method comprising:
obtaining, by the transmit end, a first data stream; processing, by the transmit end, the first data stream using an encoding neural network, to obtain a first symbol stream, (Fig. 1, ¶0047, receiving by transmitter 102 input information 108, i.e. first data stream, and process them into a series of signals. See also ¶0050-0051, 0122 )
wherein the encoding neural network comprises at least a neural network that reuses a part or all of a neural network that is the same as a decoding neural network that corresponds to the first data stream; (See at least ¶0073, encoder and decoder might share the same layer or collection of layers per Fig. 2. As such “Parameters and weight values in the network may be used for a single multiplication, as in a fully connected neural network (DNN), or they may be “tied” or replicated across multiple locations within the network”). See also ¶0123, 0161, encoder and decoder also share functions as well as training algorithm results by virtual of being trained together)
and outputting, by the transmit end, the first symbol stream. (See at least ¶0047, 0122, where the transmitter outputting the processed input signals for transmission)
Except that O’Shea does not explicitly describe the encoding neural network comprises at least a first neural network and a second neural network.
In a related field of endeavor, Van den Oord, in at least ¶0091-0093, discloses an encoder neural network structure in which the encoder neural network can include sub-neural networks, such as 410 and 414.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the encoder network of O’Shea can be implemented as a neural network made up of several sub-neural networks with a sub LSTM neural network, for example, which is configured to preserve quality of inputs with memory feature, thus can allow decoder to process output without loss of representational capacity (¶0100). Per ¶0073 of O’Shea, the entire encoder neural network can reuse an same layers of decoder, therefore, in view of Van den Oord, at least the first neural network of the encoder reuses at least part of the decoder.
As to claims 22, 42:
O’Shea in view of Van den Oord discloses all limitations of claim 21/41, O’ Shea further comprising: after outputting, by the transmit end, the first symbol stream, receiving, by the transmit end, a first weight, wherein the first weight is from the decoding neural network, and the first weight is used to train the encoding neural network. (¶0106, 109-0110, 0113, the transmission and eventual reconstruction of the input information might be a part of a close-loop training, wherein a loss function is determined from the decoder network’s output, which provides a weight update to the encoder network, i.e. training in an iterative manner)
As to claims 23, 43:
O’Shea in view of Van den Oord discloses all limitations of claim 21/41, O’ Shea further comprising: after outputting, by the transmit end, the first symbol stream, receiving, by the transmit end, a first gradient, wherein the first gradient is from the decoding neural network, and the first gradient is used to train the encoding neural network. (¶0106, 109-0110, 0113, in similar manner to claim 22, wherein a feedback is determined from the decoder network’s output, which provides gradient of the objective function, which is used to further update the encoder during training via calculating rate of change to select variations for the encoder’ parameter(s))
As to claim 24, 44:
O’Shea in view of Van den Oord discloses all limitations of claim 21/41, O’ Shea further comprising:
after outputting, by the transmit end, the first symbol stream, receiving, by the transmit end, a second function, wherein the second function is from the decoding neural network, and the second function is a loss function or a reward function; and processing, by the transmit end, the second function using the encoding neural network, to obtain a second gradient, wherein the second gradient is used to train the encoding neural network. (¶0106, 109-0113, the transmission and eventual reconstruction of the input information might be a part of a close-loop training, wherein a loss/objective function is determined from the decoder network’s output, which is used to deprive a gradient-based update by calculating a gradient of the objective function. The gradient is used to update the encoder by the variation in weights as an example)
As to claims 25, 45:
O’Shea in view of Van den Oord discloses all limitations of claim 21/41, O’ Shea discloses wherein outputting, by the transmit end, the first symbol stream comprises: performing, by the transmit end, filtering processing on the first symbol stream, to obtain a first waveform signal, wherein an out-of-band signal is filtered out from the first waveform signal; and outputting, by the transmit end, the first waveform signal. , (Fig. 1, ¶0047, receiving by transmitter 102 input information 108, i.e. first data stream, and process them into a series of signals. See also ¶0134-0135, 0122, a filter is included in the process to filter unwanted signal parts before transmission. Whether low-pass (which is part of a transmitter) or high-pass, the filter filters out the low/high signal portions before transmission )
As to claims 26, 46:
O’Shea in view of Van den Oord discloses all limitations of claim 21/41, O’Shea discloses wherein processing, by the transmit end, the first data stream using the encoding neural network, to obtain the first symbol stream comprises: performing, by the transmit end, encoding processing on the first data stream, to obtain a first channel encoding code word; and processing, by the transmit end, the first channel encoding code word using the encoding neural network, to obtain the first symbol stream. (See ¶0049, 0063, 0037, 0122-0123, the transmitter including an encoding network which, naturally, perform encoding of input data stream as part of the transmission process over the channel(s). ¶0129, the system in training determine an error rate for codework in order to update the corresponding parameter, i.e. codeword, thus involving determining a codeword).
As to claim 27:
O’Shea in view of Van den Oord discloses all limitations of claim 21, O’Shea discloses wherein the encoding neural network reusing a part or all of the neural network that is the same as the decoding neural network that corresponds to the first data stream comprises:
reusing, by the encoding neural network, a part or all of a model of the decoding neural network that corresponds to the first data stream,
a loss function of the decoding neural network that corresponds to the first data stream,
a reward function of the decoding neural network that corresponds to the first data stream,
or
a parameter of the decoding neural network that corresponds to the first data stream.
(See at least ¶0073, encoder and decoder might share the same layer or collection of layers per Fig. 2. As such “Parameters and weight values in the network may be used for a single multiplication, as in a fully connected neural network (DNN), or they may be “tied” or replicated across multiple locations within the network”). See also ¶0123, 0161, encoder and decoder also share functions as well as training algorithm results by virtual of being trained together. ¶0137, 0138 a loss function is determined and is shared to provide updates to both encoder and decoder)
As to claim 34:
O’Shea discloses:
A communication apparatus, comprising:
a processor; and a transceiver connected to the processor; wherein the processor is configured to execute program code stored in a memory, (See ¶0009, 0179, communication device with memory/processor with transceivers coupled thereto) and when the program code is executed, the communication apparatus is enabled to: obtain a first data stream; process the first data stream using an encoding neural network, to obtain a first symbol stream, (Fig. 1, ¶0047, receiving by transmitter 102 input information 108, i.e. first data stream, and process them into a series of signals. See also ¶0050-0051, 0122 )
wherein the encoding neural network reuses a part or all of as neural network that is the same as a decoding neural network that corresponds to the first data stream; (See at least ¶0073, encoder and decoder might share the same layer or collection of layers per Fig. 2. As such “Parameters and weight values in the network may be used for a single multiplication, as in a fully connected neural network (DNN), or they may be “tied” or replicated across multiple locations within the network”). See also ¶0123, 0161, encoder and decoder also share functions as well as training algorithm results by virtual of being trained together)
and output the first symbol stream. (See at least ¶0047, 0122, where the transmitter outputting the processed input signals for transmission)
As to claim 35:
O’Shea in view of Van den Oord discloses all limitations of claim 34, O’Shea discloses wherein when the program code is executed, the communication apparatus is further enabled to: receive a first weight, wherein the first weight is from the decoding neural network, and the first weight is used to train the encoding neural network. (¶0106, 109-0110, 0113, the transmission and eventual reconstruction of the input information might be a part of a close-loop training, wherein a loss function is determined from the decoder network’s output, which provides a weight update to the encoder network, i.e. training in an iterative manner)
As to claim 36:
O’Shea in view of Van den Oord discloses all limitations of claim 34, O’Shea discloses when the program code is executed, the communication apparatus is further enabled to: receive a first gradient, wherein the first gradient is from the decoding neural network, and the first gradient is used to train the encoding neural network. (¶0106, 109-0110, 0113, in similar manner to claim 22, wherein a feedback is determined from the decoder network’s output, which provides gradient of the objective function, which is used to further update the encoder during training via calculating rate of change to select variations for the encoder’ parameter(s))
As to claim 37:
O’Shea in view of Van den Oord discloses all limitations of claim 34, O’Shea discloses O’Shea discloses when the program code is executed, the communication apparatus is further enabled to:
receive a second function, wherein the second function is from the decoding neural network, and the second function is a loss function or a reward function; and process the second function using the encoding neural network, to obtain a second gradient, wherein the second gradient is used to train the encoding neural network. (¶0106, 109-0113, the transmission and eventual reconstruction of the input information might be a part of a close-loop training, wherein a loss/objective function is determined from the decoder network’s output, which is used to deprive a gradient-based update by calculating a gradient of the objective function. The gradient is used to update the encoder by the variation in weights as an example)
As to claim 38:
O’Shea in view of Van den Oord discloses all limitations of claim 34, O’Shea discloses wherein when the program code is executed, the communication apparatus is enabled to:
perform filtering processing on the first symbol stream, to obtain a first waveform signal, wherein an out-of-band signal is filtered out from the first waveform signal; and output the first waveform signal. (Fig. 1, ¶0047, receiving by transmitter 102 input information 108, i.e. first data stream, and process them into a series of signals. See also ¶0134-0135, 0122, a filter is included in the process to filter unwanted signal parts before transmission. Whether low-pass (which is part of a transmitter) or high-pass, the filter filters out the low/high signal portions before transmission )
As to claim 39:
O’Shea in view of Van den Oord discloses all limitations of claim 34, O’Shea discloses wherein when the program code is executed, the communication apparatus is enabled to: perform encoding processing on the first data stream, to obtain a first channel encoding code word; and process the first channel encoding code word using the encoding neural network, to obtain the first symbol stream. (See ¶0049, 0063, 0037, 0122-0123, the transmitter including an encoding network which, naturally, perform encoding of input data stream as part of the transmission process over the channel(s). ¶0129, the system in training determine an error rate for codework in order to update the corresponding parameter, i.e. codeword, thus involving determining a codeword).
As to claim 40:
O’Shea in view of Van den Oord discloses all limitations of claim 34, O’Shea discloses wherein the first neural network further reuse a part or all of a model of the decoding neural network, a loss function of the decoding neural network, a reward function of the decoding neural network, or a parameter of the decoding neural network.
(See discussion in claim 34, and See at least ¶0073, encoder and decoder might share the same layer or collection of layers per Fig. 2. As such “Parameters and weight values in the network may be used for a single multiplication, as in a fully connected neural network (DNN), or they may be “tied” or replicated across multiple locations within the network”). See also ¶0123, 0161, encoder and decoder also share functions as well as training algorithm results by virtual of being trained together. ¶0137, 0138 a loss function is determined and is shared to provide updates to both encoder and decoder)
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Park et al. (US 2020/0034948) - U-Net based pix architecture for transforming LR M images to sCT images. The input is encoded sequentially as a feature map of reduced spatial dimension and increased depth as it travels through the encoder layers on the left side of the network. The process is reversed as the decoder layers recover spatial information and reconstruct the output CT image. Skip connections between corresponding encoder/decoder layers, represented as grey lines at bottom here, allow shared structural features to move across the network efficiently.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 QUAN M HUA whose telephone number is (571)270-7232. The examiner can normally be reached 10:30-6:30.
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, Anthony Addy can be reached at 571-272-7795. 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.
/QUAN M HUA/Primary Examiner, Art Unit 2645
Prosecution Insights