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 .
Response to Amendment
The amendments overcome the objections to the drawings and the specification. The amendments overcome the previously applied rejections based on 35 USC section 112b. The amendments and corresponding explanations overcome the rejections based on 35 USC section 112a of the independent claims. The amendments overcome the previously applied prior art rejections. In particular, Banerjee taught the concept of monitoring packets lost and avoiding retransmission until a threshold limit of packets lost is reached, as shown in the last office action, but Banerjee did not anticipate or suggest the amended claim language specified in the final limitation where a transport layer delivers to a receiver application both data payloads of received packets and substitute data generated for any unreceived packets. Banerjee had been previously applied to claim 4 but the current claim amendment is much more specific than claim 4. The Examiner did not find any prior art that anticipated this feature or showed this feature to be obvious after an updated search.
Response to Arguments
Applicant's arguments filed on 12/3/2025 with respect to the written description rejection that had been applied to claim 4, 11, and 18 in the last office action are not persuasive. The applicant examine is not rejecting the applicant’s claims based on the enablement requirement of 35 USC section 112a. The examiner is rejecting the claims based on the written description requirement of 35 USC section 112a based on the guidance given in sections 2162 and 2161.01(I) of the MPEP. The applicant has not disclosed any algorithm for generating substitute data, the applicant did not incorporate any techniques for doing such by reference, and the applicant has not provided any evidence that such techniques are well known.
The applicant argues the following specifically:
It is respectfully submitted that performing such data substitution "in a manner that minimizes approximation error or bias of the substitute data" is enabled because a person of ordinary skill in the art as of the September 5, 2023 filing date would have been thoroughly familiar with well-established techniques for approximating missing gradient values in training systems for artificial neural networks. Numerous gradient-approximation strategies-such as using the most recently received gradient for the same coordinates, momentum-based extrapolation, zero-fill with residual correction (error feedback), statistical reconstruction from quantization parameters, and sparsification-mask-based interpolation-were routinely employed to ensure that reconstructed gradient values preserve the structure and accuracy of the overall gradient while minimizing approximation error or bias. These techniques were widely described in the literature and commonly implemented in distributed SGD (Stochastic Gradient Descent), quantized-gradient methods, and compressed communication frameworks. See, e.g., Frank Seide et al., "1-bit Stochastic Gradient Descent and Its Application to Data-Parallel Distributed Training of Speech DNNs," in Proceedings of INTERSPEECH 2014 (2014). Because the specification teaches that substitute data may be generated for lost gradient coordinates using a strategy that minimizes approximation error or bias (paragraph [0047]), and because a skilled artisan would have known multiple well-established ways to generate such substitute gradients without undue experimentation, the newly added limitations are fully enabled under 35 U.S.C. §112(a).
This argument is not persuasive for a number of reasons. The applicant’s argument that “a skilled artisan would have known multiple well-established ways to generate such substitute gradients” ignores that the applicant did not disclose anything about what learning model they are using (see paragraph 2). A skilled artisan would recognize that there are hundreds of techniques for implementing machine learning, each with their own ways for implementing gradients. A skilled artisan would look to the applicant’s disclosure and see that the applicant has not disclosed anything about how the applicant defines a gradient in a way that allows a skilled artisan to infer how a substitute gradient would be generated.
The applicant’s reference to a document by “Frank Seide” is not useful. First, the applicant did not supply a copy of this document in accordance with 37 CFR 1.97 or 1.98 and section 609.05(C) of the MPEP. Second, the applicant did not incorporate this document by reference under 37 CFR 1.57 so there appears to be no suggestion in the originally filed disclosure to go and obtain the Seide document in order to figure out what the applicant’s invention is doing to fabricate lost gradient data. Third, the applicant did not provide any description of what part of this document supposedly shows generating lost gradients and the Examiner did not see anything relevant in an online copy. If the applicant wants to insist this document is useful, they need to submit it and explain how it shows generating gradients in the context of replacing those lost during transmission.
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.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
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 of carrying out his invention.
Claims 1, 2, 7-9, 14-16, and 21-24 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains 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, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Written Description Issue #1
Claims 1, 8, and 15 recite the following limitation:
comparing the running count of lost-data size with the data loss threshold to determine whether retransmission is to be requested; and
delivering, by the transport layer, to the receiver application, (i) data payloads of the received packets and (ii) substitute data generated for any unreceived packet when the running count does not exceed the data-loss threshold, such that an amount of data replaced by the substitute data remains within the data loss threshold.
As shown in Figure 4, the applicant disclosed that the delivery claimed only happens if the running count in the comparison step is less than the data loss threshold. The applicant’s claim covers the delivery happening regardless of the comparison.
Written Description Issue #2
Claims 1, 8, and 15 recite the following limitation:
delivering, by the transport layer, to the receiver application, (i) data payloads of the received packets and (ii) substitute data generated for any unreceived packet when the running count does not exceed the data-loss threshold, such that an amount of data replaced by the substitute data remains within the data loss threshold.
Claim 22-24 recite the following limitation:
wherein the data packets contain gradients for training an artificial neural network, and the substitute data is generated to replace gradients in any unreceived packet and in a manner that minimizes approximation error or bias of the substitute data.
Section 2162 of the MPEP states:
To obtain a valid patent, a patent application must contain a full and clear description of the invention for which a patent is sought in the manner prescribed by 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph. The requirement for an adequate written description ensures that the public receives something in return for the exclusionary rights that are granted to the inventor by a patent. The grant of a patent helps to foster and enhance the development and disclosure of new ideas and the advancement of scientific knowledge. Upon the grant of a patent in the U.S., information contained in the patent becomes a part of the information available to the public for further research and development, subject only to the patentee’s right to exclude others during the life of the patent.
Section 2163.03(V) of the MPEP states:
V. ORIGINAL CLAIM NOT SUFFICIENTLY DESCRIBED
While there is a presumption that an adequate written description of the claimed invention is present in the specification as filed. In re Wertheim, 541 F.2d 257, 262, 191 USPQ 90, 96 (CCPA 1976), a question as to whether a specification provides an adequate written description may arise in the context of an original claim. An original claim may lack written description support when (1) the claim defines the invention in functional language specifying a desired result but the disclosure fails to sufficiently identify how the function is performed or the result is achieved or (2) a broad genus claim is presented but the disclosure only describes a narrow species with no evidence that the genus is contemplated. See Ariad Pharms., Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1349-50 (Fed. Cir. 2010) (en banc). The written description requirement is not necessarily met when the claim language appears in ipsis verbis in the specification. "Even if a claim is supported by the specification, the language of the specification, to the extent possible, must describe the claimed invention so that one skilled in the art can recognize what is claimed. The appearance of mere indistinct words in a specification or a claim, even an original claim, does not necessarily satisfy that requirement."Enzo Biochem, Inc. v. Gen-Probe, Inc., 323 F.3d 956, 968, 63 USPQ2d 1609, 1616 (Fed. Cir. 2002).
The applicant’s quality of the artificial neural network is directly dependent upon the gradients used for training the artificial neural network. Generating replacements gradients in a manner which minimizes approximation error or bias would require some sort of solution that is able to infer what the gradient values should have been in the context of the neural network. The applicant has not disclosed anything about how this inference is accomplished. Those of ordinary skill in the art would recognize that training a learning model with random gradients will produce an inferior learning model. Those of skilled in the art would expect a description of how the applicant is actually attempting to minimize approximation error and bias. The applicant has failed to do this. The applicant has not even described any specific learning techniques that they apply. Those skilled in the art would recognize that machine learning, in the broad context described in paragraph 2 of the applicant’s disclosure, is a broad term that does not cover any specific machine learning techniques, but rather covers a myriad of hundreds of techniques that each have their own unique sophistication with how they implement training and gradients. The applicant’s disclosure trivializes what would be recognized in the art as a complex problem.
There is no description provided by the applicant of how to create substitutes for lost data. Paragraphs 26 states that rows of a dataset can be “synthetically created” but does not describe what this synthetically created data is or how it is actually created. Paragraph 37 states that server application 304 can create “synthetic substitutes” for data in lost data frames. Paragraphs 37 references “using a strategy that minimizes approximation error or bias” but does not provide any description of how such a strategy is implemented. In other words, paragraph 37 describes the benefits of having a strategy but does not actually describe what the strategy that provides those benefits is. Paragraph 38 states that the transport layer may handle data substitution instead of the server application but does not provide any description of the creation of “synthetic data”. Paragraphs 39 and 50 further reference using “synthetic data” but do not provide any description of how such data is created. For the reasons explained with respect to the independent claims, the applicant has failed to provide a description of generating substitute data that meets the standards of written description specified in 35 USC section 112(a).
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.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1, 2, 7-9, 14-16, and 21-24 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 1, 8, and 15 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being incomplete for omitting essential steps, such omission amounting to a gap between the steps. See MPEP § 2172.01. The omitted steps are: The applicant covers delivering substitute data generated for any unreceived packets but the applicant does not actually claim the step of generating such data. The data cannot be delivered if it is not first generated.
Conclusion
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 DOUGLAS B BLAIR whose telephone number is (571)272-3893. The examiner can normally be reached Monday-Friday 9am-5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Glenton Burgess can be reached at 571-272-3949. 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.
/DOUGLAS B BLAIR/Primary Examiner, Art Unit 2454
Prosecution Insights