CTNF 18/875,141 CTNF 95839 Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. Claim Rejections - 35 USC § 103 07-06 AIA 15-10-15 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. 07-20-aia AIA 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. 07-23-aia AIA 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. 07-21-aia AIA Claim s 1, 2, 3, 9, 10, 19 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Srikotr (T. Srikotr and K. Mano, "Predictive Vector Quantized Variational AutoEncoder for Spectral Envelope Quantization," 2020 International Conference on Electronics, Information, and Communication (ICEIC), Barcelona, Spain, 2020, pp. 1-4, doi: 10.1109/ICEIC49074.2020.9051233)and Garon (US 20190149625 A1) . With respect to claims 1, 19 and 20 Srikotr teaches (claim 1 ). (Original) A computing system comprising: at least one hardware processor; at least one memory coupled to the at least one hardware processor; and one or more computer-readable storage media comprising computer-executable instructions that, when executed, cause the computing system to perform operations comprising (Srikotr , see section IV where experimentation and results are discussed, which inherently allude to having computer program, a non-transitory computer readable storage medium having program instructions embodied therewith, a system comprising a memory device for storing program code; and a processor device operatively coupled to the memory device for running the program code.) : (claim 19 ) A method, implemented in a computing system comprising at least one hardware processor and at least one memory coupled to the at least one hardware processor, the method comprising (Srikotr , see section IV where experimentation and results are discussed, which inherently allude to having computer program, a non-transitory computer readable storage medium having program instructions embodied therewith, a system comprising a memory device for storing program code; and a processor device operatively coupled to the memory device for running the program code.) : (claim 20 ) One or more computer-readable storage media comprising: computer-executable instructions that, when executed by a computing system comprising at least one hardware processor and at least one memory coupled to the at least one hardware processor, cause the computing system to (Srikotr , see section IV where experimentation and results are discussed, which inherently allude to having computer program, a non-transitory computer readable storage medium having program instructions embodied therewith, a system comprising a memory device for storing program code; and a processor device operatively coupled to the memory device for running the program code.) extracting one or more latent features from a frame of an input signal using an encoder to provide extracted one or more latent features (Srikotr ¶Page 2, Section Ill.: In the encoding process, the input data is fed into the encoder network to produce the z-latent ze(x). It is reshaped in N sub-vector (reshaped ze(x)) corresponding to the designed embedding space"; figure 2: "VQ-VAE Encoder'', "ze(x)");) ; determining a prediction of the one or more latent features using reconstructed latent features for a plurality of prior frames (Srikotr ¶Page 2, Section Ill.: "the quantized z-latent zq(x) is utilized as the input to the encoder predictor network. The predicted z-latent zq(x) is appropriated to subtract with the next ze(x) and add to the next zq(x). "; figure 2: "Encoder Predictor");) extracting a residual-like feature from the extracted one or more latent features and the prediction (Srikotr ¶Page 2, Section Ill col2 para1: The predicted z-latent zq(x) is appropriated to subtract with the next ze(x)"; figure 2: difference between the z-latent and the predicted z-latent is calculated as a residual) Srikotr does not explicitly disclose however Garon teaches and sending the residual-like feature, or data sufficient to reconstitute the residual-like feature, to a client (Garon ¶ [0002] In general, one innovative aspect of the subject matter described in this specification can be embodied in a method that includes identifying, by one or more servers, an opportunity to transmit a digital component to a client device that is identified by a given unique identifier). It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr to include the transmission of Garon in order to optimize bandwidth and increase communication efficiency. With respect to claim 2 Srikotr teaches wherein the input signal comprises audio data (Srikotr ¶Page 1, Section I.: to decrease speech information for saving communication bandwidth and reduce complexity'') . With respect to claim 3 Srikotr teaches , wherein the extracting comprises the use of at least one convolution layer (Srikotr ¶Page 2, Section IV. The encoder network was implemented as the two stride convolutional layers) With respect to claim 9 Srikotr teaches wherein the encoder comprises a plurality of convolution layers (Srikotr ¶Page 3, Section IV.B.: "Speech database", "The encoder network was implemented as the two stride convolutional layers") ) . With respect to claim 10 Srikotr teaches wherein the determining a prediction comprises processing the reconstructed latent features for the plurality of prior frames using a plurality of convolution layers (Srikotr ¶¶Page 3, Section IV.B.: "Speech database", "The encoder network was implemented as the two stride convolutional layers",¶P2, SecIV B: The decoder network was implemented as one transposed convolutional layer attached to the two residual networks, which transposed the convolutional layers.) . 07-21-aia AIA Claim s 4, 5 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Srikotr, Garon and Iskander (US 20140323823 A1) . With respect to claim 4 Srikotr Garon do not explicitly disclose however, Iskander teaches wherein input signal comprises time-frequency spectrum data (Iskander ¶ [0055] Then a windowed STFT operation is applied and the spectra corresponding to different windows are averaged to increase the SNR) It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon to include time-frequency spectrum of Iskander in order to reveal temporal evolution of specific frequency information. With respect to claim 5 Srikotr Garon do not explicitly disclose however, Iskander teaches wherein the time-frequency spectrum data is obtained using a short-time Fourier transform of a time window of the input signal (Iskander ¶ [0055] Then a windowed STFT operation is applied and the spectra corresponding to different windows are averaged to increase the SNR) Examiner Note: STFT is windowed and in the frequency domain maps to a bank of filters that the examiner maps to groups in the frequency domain. It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon to include time-frequency spectrum of Iskander in order to reveal temporal evolution of specific frequency information. With respect to claim 12 Srikotr, Garon do not explicitly disclose however, Iskander teaches wherein a given group of the plurality of groups comprises a plurality of frequencies (Iskander ¶ [0055] Then a windowed STFT operation is applied and the spectra corresponding to different windows are averaged to increase the SNR). It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon to include time-frequency spectrum of Iskander in order to reveal temporal evolution of specific frequency information . 07-21-aia AIA Claim s 11 are rejected under 35 U.S.C. 103 as being unpatentable over Srikotr, Garon, Iskander and Fesseler (US 5303346 A) . With respect to claim 11 Srikotr, Garon do not explicitly disclose however, Iskander teaches splitting the residual-like feature into a plurality of groups along a channel dimension wherein a given group of the plurality of groups comprises a plurality of frequencies (Iskander ¶ [0055] Then a windowed STFT operation is applied and the spectra corresponding to different windows are averaged to increase the SNR). It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon to include time-frequency spectrum of Iskander in order to reveal temporal evolution of specific frequency information. None of Srikotr, Garon and Iskander explicitly disclose however, Fesseler teaches and separately quantizing groups of the plurality of groups (Col1ll39-42: These magnitude groups are then routed, according to frequency range, to five different quantizers which quantize nonuniformly, approximating the logarithmic loudness perception.) It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon in view of time-frequency spectrum of Iskander to include separate quantization of groups of Fesseler in order to preserve accuracy for non-uniform data in each group . 07-21-aia AIA Claim s 6-8 are rejected under 35 U.S.C. 103 as being unpatentable over Srikotr, Garon, Iskander and Jiang (Jiang, Xue, et al. "End-to-end neural speech coding for real-time communications." ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2022.) . With respect to claim 6 none of Srikotr, Garon and Iskander explicitly disclose however, Jiang teaches the operations further comprising applying amplitude compression to the time-frequency spectrum data (Jiang ¶p2Col1Sec2.1. Overview The TFNet-based codec takes the time-frequency spectrum (20ms window with a 5ms hop length) as input with a power-law compression on the amplitude before feeding into the net-work. As the dynamic range of speech is high due to harmonics, the compression performs as a kind of input normalization so that the importances of different frequencies are balanced and the training is more stable.) It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon in view of time-frequency spectrum of Iskander to include amplitude compression of Jiang in order to compress dynamic range With respect to claim 7 none of Srikotr, Garon and Iskander explicitly disclose however, Jiang teaches wherein the amplitude compression is applied using a value determined during training of the encoder (Jiang ¶p2Col1Sec2.1. Overview The TFNet-based codec takes the time-frequency spectrum (20ms window with a 5ms hop length) as input with a power-law compression on the amplitude before feeding into the net-work. As the dynamic range of speech is high due to harmonics, the compression performs as a kind of input normalization so that the importances of different frequencies are balanced and the training is more stable .) It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon in view of time-frequency spectrum of Iskander to include amplitude compression of Jiang in order to compress dynamic range With respect to claim 8 none of Srikotr, Garon and Iskander explicitly disclose however, Jiang teaches wherein the value differs for different encoding bitrates (Jiang ¶p2Col1Sec2.1. Overview The TFNet-based codec takes the time-frequency spectrum (20ms window with a 5ms hop length) as input with a power-law compression on the amplitude before feeding into the net-work. As the dynamic range of speech is high due to harmonics, the compression performs as a kind of input normalization so that the importances of different frequencies are balanced and the training is more stable .) It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify latent feature extraction of Srikotr in view of transmission of Garon in view of time-frequency spectrum of Iskander to include amplitude compression of Jiang in order to compress dynamic range Allowable Subject Matter 07-43 Claims 13, 14- 18 are objected to as being dependent upon a rejected base claim, but would be allowable if written in independent form including all of the limitations of the base claim and any intervening claims. Claim 13 recites wherein the channels are quantized using different codebooks, the operations further comprising, during training of the encoder: for a set of input training data used during training of the encoder, randomly selecting a group of the plurality of groups, wherein groups are associated with sets of progressively higher bitrates; and during training of the encoder using the set of input training data, using only the selected group of the plurality of groups and groups of the plurality of groups associated with lower bitrates than the selected group. The closest teachings come from Jiang who teaches “Sec 2.4. Vector Quantization: The vector quantizer discretizes the learned features in encoding with a set of learnable codebooks according to the target bitrate. Before quantization, the features after encoding XS ∈ RT×1×C are reduced to XQ ∈ RT×1×C′ through a 1 × 1 convolution (C′ < C). We take a group quantization by splitting channels C′ into N groups and coding each group by an independent codebook. Let S denote the number of codewords in each codebook and K = C′/N the dimension of each codeword. In the proposed scheme, a window length of 20ms and hop length of 5ms is adopted for STFT and thus the bitrate is given by N × log2S/5 kbps. For 6kbps, C′, N, S and K are set to 120, 3, 1024, and 40, respectively. The codebooks are learned with exponential moving average, following that in [11]. The quantized features X^Q ∈ RT×1×C′ are enlarged to the shape T×1×C before feeding into the temporal filtering blocks in decoding.” However, none of the prior art of record including Jiang teach the limitation as stated above specifically the underlined as shown including all supporting limitations thereof. Therefore claim 13 is allowable. Claim 14 recites quantizing the residual-like feature, the quantizing comprising: for the frame, determining a distance between the residual-like feature and a codeword of a codebook used for vector quantization of the residual-like feature; and determining a probability of selecting the codeword at least in part using the distance. The closest teachings come from Jiang who teaches “Sec 2.4. Vector Quantization: The vector quantizer discretizes the learned features in encoding with a set of learnable codebooks according to the target bitrate. Before quantization, the features after encoding XS ∈ RT×1×C are reduced to XQ ∈ RT×1×C′ through a 1 × 1 convolution (C′ < C). We take a group quantization by splitting channels C′ into N groups and coding each group by an independent codebook. Let S denote the number of codewords in each codebook and K = C′/N the dimension of each codeword. In the proposed scheme, a window length of 20ms and hop length of 5ms is adopted for STFT and thus the bitrate is given by N × log2S/5 kbps. For 6kbps, C′, N, S and K are set to 120, 3, 1024, and 40, respectively. The codebooks are learned with exponential moving average, following that in [11]. The quantized features X^Q ∈ RT×1×C′ are enlarged to the shape T×1×C before feeding into the temporal filtering blocks in decoding.” However, none of the prior art of record including Jiang teach the limitation as stated above specifically the underlined as shown including all supporting limitations thereof. Therefore claim 14 is allowable. Claim 15-16 are allowable because of dependency on 14. Claim 17 recites wherein the residual-like feature, or the data sufficient to reconstitute the residual-like feature, is sent as part of a bitstream having a rate, the operations further comprising: during training of the encoder, determining a bitrate for training input data, the determining a bitrate comprising determining a difference between a target bitrate and an entropy of probabilities of selecting particular codewords of a codebook for frames of the training input data. The closest teachings come from Jiang who teaches “Sec 2.4. Vector Quantization: The vector quantizer discretizes the learned features in encoding with a set of learnable codebooks according to the target bitrate. Before quantization, the features after encoding XS ∈ RT×1×C are reduced to XQ ∈ RT×1×C′ through a 1 × 1 convolution (C′ < C). We take a group quantization by splitting channels C′ into N groups and coding each group by an independent codebook. Let S denote the number of codewords in each codebook and K = C′/N the dimension of each codeword. In the proposed scheme, a window length of 20ms and hop length of 5ms is adopted for STFT and thus the bitrate is given by N × log2S/5 kbps. For 6kbps, C′, N, S and K are set to 120, 3, 1024, and 40, respectively. The codebooks are learned with exponential moving average, following that in [11]. The quantized features X^Q ∈ RT×1×C′ are enlarged to the shape T×1×C before feeding into the temporal filtering blocks in decoding, ¶(Jiang ¶p2Col1Sec2.1. Overview The TFNet-based codec takes the time-frequency spectrum (20ms window with a 5ms hop length) as input with a power-law compression on the amplitude before feeding into the net-work. As the dynamic range of speech is high due to harmonics, the compression performs as a kind of input normalization so that the importances of different frequencies are balanced and the training is more stable.)” However, none of the prior art of record including Jiang teach the limitation as stated above specifically the underlined as shown including all supporting limitations thereof. Therefore claim 17 is allowable and 18 is allowable because of its dependency. Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to ATHAR N PASHA whose telephone number is (408)918-7675. The examiner can normally be reached on Monday-Thursday Alternate Fridays, 7:30-4:30 PT. 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, Daniel Washburn can be reached on (571)272-5551. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000. /ATHAR N PASHA/Primary Examiner, Art Unit 2657 Application/Control Number: 18/875,141 Page 2 Art Unit: 2657 Application/Control Number: 18/875,141 Page 3 Art Unit: 2657 Application/Control Number: 18/875,141 Page 4 Art Unit: 2657 Application/Control Number: 18/875,141 Page 5 Art Unit: 2657 Application/Control Number: 18/875,141 Page 6 Art Unit: 2657 Application/Control Number: 18/875,141 Page 7 Art Unit: 2657 Application/Control Number: 18/875,141 Page 8 Art Unit: 2657 Application/Control Number: 18/875,141 Page 9 Art Unit: 2657 Application/Control Number: 18/875,141 Page 10 Art Unit: 2657 Application/Control Number: 18/875,141 Page 11 Art Unit: 2657
Prosecution Insights