DETAILED ACTION
This Office Action is sent in response to Applicant’s Communication received 11/28/2023 for application number 18/521,286.
Claims 1-20 are pending.
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 .
Claim Rejections - 35 USC § 102
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:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claim(s) 1-4, 6, 8-11, and 13 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Williams et al., Gradient-Based Learning Algorithms for Recurrent Networks and Their Computational Complexity (NPL [U], see Notice of References Cited).
In reference to claim 1, Williams discloses a method of training a neural network, the neural network comprising operators defined by weights associated with nodes in the neural network (method for training RNN comprising nodes and weights, pages 447, 433-35), comprising: partitioning an ordered sequence of training sets over a recurrent window of training sets into a sequence of training set batches, each training set comprising one or more input features, the sequence of training set batches comprising a plurality of initial training set batches followed by a final training set batch (training data is partitioned into a plurality of time step, which are batches, with a final batch at
t
1
, pages 447-49, fig. 4); for at least one of the initial training set batches: executing the neural network to compute one or more inferences based on input features of training sets in at least one of the initial training set batches (forward passes are performed with initial timestep data to get output, pages 447-49, fig. 4); and computing an updated state of a persistent history buffer, wherein the neural network further computes the inferences in the at least one of the initial training set batches based on a state of the persistent history buffer, the persistent history buffer updated from execution of the neural network based on one or more preceding initial training set batches (history buffer stores input and output for each forward pass for each timestep, pages 447-49, fig. 4); for at least the final training set batch: executing the neural network to compute one or more inferences based on input features of the final training set batch, wherein the neural network further computes the one or more inferences in the final training set batch based on the updated state of the persistent history buffer computed from executing the neural network based on the input features of the at least one of the initial training set batches; and updating the weights of the neural network using a gradient descent backpropagation based on the computed one or more inferences based on input features of the final training set batch and ground truth observations for the input features of the final training set batch (at final timestep of batch, backpropagation is performed using values from history buffer to calculate loss, and weights are updated, pages 447-49, fig. 4).
In reference to claim 2, Williams discloses the method of claim 1, and further comprising: updating the weights of the neural network using the gradient descent backpropagation based on the computed one or more inferences based on input features of the at least one of the initial training set batches and ground truth observations for the input features of the at least one of the initial training set batches prior to executing the neural network to compute one or more inferences based on input features of the final training set batch (at final timestep, backpropagation is performed using values from history buffer that were calculated from each previous timestep / batch to calculate loss, and weights are updated, pages 447-49, fig. 4).
In reference to claim 3, Williams discloses the method of claim 1, wherein the weights are unchanged through execution of the neural network for prior to updating the weights based on the computed one or more inferences based on input features of the final training set batch and ground truth observations for the input features of the final training set batch (see pages 447-49 and fig. 4: the weights are not updated until the backward pass at the final timestep).
In reference to claim 4, Williams discloses the method of claim 1, wherein an accumulation of inferences computed in a preceding training batch are maintained in a history buffer, and the method further comprises: resetting the history buffer; and accumulating inferences computed in a training batch in the history buffer (after backpropagation, network is reinitialized, including buffer, and training process is repeated, page 449).
In reference to claim 6, Williams discloses the method of claim 1, wherein the neural network comprises a recurrent neural network (RNN) (pages 433-35).
In reference to claim 8, this claim is directed to a device associated with the method claimed in claim 1 and is therefore rejected under a similar rationale.
In reference to claim 9, this claim is directed to a device associated with the method claimed in claim 2 and is therefore rejected under a similar rationale.
In reference to claim 10, this claim is directed to a device associated with the method claimed in claim 3 and is therefore rejected under a similar rationale.
In reference to claim 11, this claim is directed to a device associated with the method claimed in claim 4 and is therefore rejected under a similar rationale.
In reference to claim 13, this claim is directed to a device associated with the method claimed in claim 6 and is therefore rejected under a similar rationale.
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.
Claim(s) 5, 7, 12, 14, and 15-19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Williams et al., Gradient-Based Learning Algorithms for Recurrent Networks and Their Computational Complexity (NPL [U], see Notice of References Cited) in view of Weinzaepfel (US 2020/0160065 A1).
In reference to claim 5, Williams does not explicitly teach the method of claim 1, wherein the input features of the training set batches comprise features of images.
Weinzaepfel teaches the method of claim 1, wherein the input features of the training set batches comprise features of images (sequences of images are used as training data for RNN, para. 0032-34).
It would have been obvious to one of ordinary skill in art, having the teachings of Williams and Weinzaepfel before the earliest effective filing date, to modify the training data of Williams to include the images of Weinzaepfel.
One of ordinary skill in the art would have been motivated to modify the training data of Williams to include the images of Weinzaepfel because it would allow the RNN of Williams to process more types of data and tasks, like image entity recognition (Weinzaepfel, para. 0004-07).
In reference to claim 7, Williams not explicitly teach the method of claim 1, and further comprising: warping an accumulation of inferences from execution of the neural network based on the one or more preceding initial training set batches; and executing the neural network to compute the inferences in at least one training batch of the initial training set batches based on the warped accumulation of inferences.
Weinzaepfel teaches the method of claim 1, and further comprising: warping an accumulation of inferences from execution of the neural network based on the one or more preceding initial training set batches; and executing the neural network to compute the inferences in at least one training batch of the initial training set batches based on the warped accumulation of inferences (sequences of image frames are warped to shift and align images, and the warped sequence is used for inferences, para. 0046-63).
It would have been obvious to one of ordinary skill in art, having the teachings of Williams and Weinzaepfel before the earliest effective filing date, to modify the training data of Williams to include the warping of Weinzaepfel.
One of ordinary skill in the art would have been motivated to modify the training data of Williams to include the warping of Weinzaepfel because it would allow the RNN of Williams to process more types of data and tasks, like image entity recognition, while providing better performance (Weinzaepfel, para. 0004-07, 0125).
In reference to claim 12, this claim is directed to a device associated with the method claimed in claim 5 and is therefore rejected under a similar rationale.
In reference to claim 14, this claim is directed to a device associated with the method claimed in claim 7 and is therefore rejected under a similar rationale.
In reference to claim 15, Williams teaches a method of training a recurrent neural network, (method for training RNN, pages 447, 433-35), comprising: receiving a windowed sequence of input … in an input layer of a recurrent neural network; reading a sequence of output … from an output layer of the recurrent neural network corresponding to the windowed sequence of input … (training data is divided into epochs and timesteps, and training data is input to RNN to get output, pages 447-49, fig. 4); providing a sequence of ground truths corresponding to the windowed sequence of input … and representing a desired output; training the recurrent neural network to predict the provided sequence of ground truths based on the received windowed sequence of input tensors by using backpropagation to adjust a weight of one or more activation functions linking one or more nodes of one or more layers of the recurrent neural network based on a difference between the sequence of output … and the ground truths (backpropagation is performed with desired output to calculate loss, and weights of nodes of the RNN are updated, pages 447-49, fig. 4); and maintaining a persistent history buffer comprising one or more states of the recurrent neural network across multiple groups of input …, and providing data from the persistent history buffer to the recurrent neural network as an input while training the recurrent neural network (history buffer stores input, output, and desired output for each forward pass for across epochs, or groups, of timesteps, and data in history buffer is used during backpropagation for loss calculation to update weights, pages 447-49, fig. 4)
However, Williams does not explicitly teach tensors.
Weinzaepfel teaches tensors (input and output are tensors, para. 0036-43, and particularly para. 0040).
It would have been obvious to one of ordinary skill in art, having the teachings of Williams and Weinzaepfel before the earliest effective filing date, to modify the training of Williams to include the tensors of Weinzaepfel.
One of ordinary skill in the art would have been motivated to modify the training of Williams to include the tensors of Weinzaepfel because it would allow the RNN of Williams to process more types of data and tasks, like image entity recognition (Weinzaepfel, para. 0004-07).
In reference to claim 16, Williams teaches the method of training a recurrent neural network of claim 15, further comprising resetting the persistent history buffer at end of a sequence (after backpropagation, network is reinitialized, including buffer, and training process is repeated, page 449).
In reference to claim 17, Williams and Weinzaepfel teach the method of training a recurrent neural network of claim 15, wherein the groups of input tensors comprise windows of input tensors (Williams teaches the training data are partitioned into a plurality of time step, which are windows of data, pages 447-49, fig. 4; Weinzaepfel teaches tensors over time, para. 0028-45).
In reference to claim 18, Williams and Weinzaepfel teach the method of training a recurrent neural network of claim 15, wherein the groups of input tensors combine training batches of input tensors (epochs are across multiple groups of input, pages 447-49, fig. 4; Weinzaepfel teaches tensors over time, para. 0028-45).
In reference to claim 19, Williams teaches the method of training a recurrent neural network of claim 15, wherein training the recurrent neural network comprises performing backpropagation at an end of a training batch (at final timestep of batch, backpropagation is performed using values from history buffer to calculate loss, and weights are updated, pages 447-49, fig. 4).
Claim(s) 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Williams et al., Gradient-Based Learning Algorithms for Recurrent Networks and Their Computational Complexity (NPL [U], see Notice of References Cited) in view of Weinzaepfel (US 2020/0160065 A1) as applied to claim 15 above, and in further view of Polak et al. (US 2019/0294881 A1).
In reference to claim 20, Williams and Weinzaepfel do not explicitly teach the method of training a recurrent neural network of claim 15, further comprising randomizing at least one of a temporal position of a training window within a training sequence and an order of training sequences in a training epoch to reduce overfitting during training.
Polak teaches the method of training a recurrent neural network of claim 15, further comprising randomizing at least one of a temporal position of a training window within a training sequence and an order of training sequences in a training epoch to reduce overfitting during training (extracted window is randomly selected at each iteration, which is randomizing training sequences, para. 0102).
It would have been obvious to one of ordinary skill in art, having the teachings of Williams. Weinzaepfel, and Polak before the earliest effective filing date, to modify the training of Williams to include the randomization of Polak.
One of ordinary skill in the art would have been motivated to modify the training of Williams to include the randomization of Polak because it can help determine an optimal window size (Polak, para. 0102).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Andrew T. Chiusano whose telephone number is (571)272-5231. The examiner can normally be reached M-F, 10am-6pm.
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, Tamara Kyle can be reached at 571-272-4241. 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.
/ANDREW T CHIUSANO/Primary Examiner, Art Unit 2144
Prosecution Insights