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 .
Detailed Action
The following action is in response to the communication(s) received on 12/19/2025.
As of the claims filed 12/19/2025:
Claims 1-4 and 7-9 have been amended.
Claims 5 and 6 have been canceled.
Claims 1-4 and 7-9 are now pending.
Claims 1, 8, and 9 are independent claims.
Response to Arguments
Applicant’s arguments filed 12/19/2025 have been fully considered, but are not fully persuasive.
The objection to the title has been withdrawn in view of the amendment.
With respect to the art rejections under 35 USC § 101:
Applicant asserts that the instant claims are eligible for the same reasons as Example 39, as they are directed to a training mechanism and thus only based on abstract ideas (p.10 ¶2-3). Examiner respectfully disagrees, as the claims as currently recited are focused on the abstract ideas without sufficient details to the architecture of the training. For example, the amended claim “sampling the temperature parameter…” is required to be performed “during learning”, which is an abstract idea and not a technology of training a neural network. Thus, the claims do remain reciting abstract ideas and dissimilar to Example 39.
Applicant further asserts that the claims are directed to an improvement in the training of a conditional neural network (p.11 ¶2). Similar to above, the claims are merely generally linked to the training phase instead of being directed to a particular training step of the conditional neural network. Thus, the claims do not recite an integration of the abstract ideas into a practical application.
Applicant further asserts that the updating of the neural parameters using the error function is non-conventional and thus significantly more than the abstract ideas (p.12). However, the error function calculations themselves are an abstract idea and thus are not determined whether it is significantly more. Additionally, the location between the output from the feature and target information are merely generally linked to the abstract idea of error calculation and not a particular training step, as mentioned above. Thus, the amended limitations do not recite significantly more than the abstract ideas.
With respect to the art rejections under 35 USC § 102 and § 103:
Applicant asserts that the amended limitations, specifically the error loss and the uniform distribution sampling, are not taught by Wang. Examiner respectfully disagrees, as Wang does teach these limitations: Wang, p.6 ¶2, where the Gumbel distribution g is sampled by drawing u ~Uniform and thus corresponds to the uniform distribution. Additionally, (Wang, Algorithm 1 line 12, [p.6 ¶6]), where updating D corresponds to updating the neural network parameters; the loss function in line 6 corresponds to the error function between the output (yc) and the target information (y)).
Applicant further asserts that the dependent claims 2-4 are allowable by virtue of dependency of their parent claim. Examiner respectfully submits that claim 1 remains rejected at least by virtue of dependency of the rejected parent claim 1.
Claims 5 and 6 have been canceled. Thus, the rejections for claims 5 and 6 have been withdrawn.
Claim Rejections - 35 USC § 112
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 1-4 and 7 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.
Claim 1 recites the limitation "the gradient of the error function". There is insufficient antecedent basis for this limitation in the claim.
Claims 2-4 and 7 are rejected by at least by virtue of dependency to their parent claim.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-4 and 7-9 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 recites a system, thus a machine, one of the four statutory categories of patentable subject matter (Step 1). However, Claim 1 further recites:
generate a condition vector from training data, which is an evaluation or judgement that can be performed in the human mind;
connect the condition vector to a feature quantity of the training data, which is an evaluation or judgement that can be performed in the human mind;
apply a softmax activation with a temperature parameter T to generate the condition vector, the condition vector being a one-hot vector, which is an evaluation or judgement that can be performed in the human mind;
during learning, sampling the temperature parameter T for the softmax from a uniform distribution U(Tmin, Tmax)…, independently for each training step, which is an evaluation or judgement that can be performed in the human mind;
connecting the condition vector to the feature quantity of an input information by concatenation in a channel direction to form a combined feature which is an evaluation or judgement that can be performed in the human mind;
optimize a parameter of the neural network by using the feature quantity to which the condition vector is connected, and compute an error function between an output decoded from the combined feature and the target information and update the neural network parameters by back-propagation using the gradient of the error function, which is a mathematical concept.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
comprising: at least one memory that is configured to store information and instructions; and at least one processor that is configured to execute the instructions, the instructions causing the at least one processor to:…, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application;
by a temperature sampling unit, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application;
obtaining, from target information in the training data, a feature vector, which is merely an insignificant extra-solution activity of data gathering, which by MPEP 2106.05(g) cannot integrate an abstract idea into a practical application.
that are inputted to a neural network, which merely specifies the particular field of use or particular technological environment in which the abstract idea is to be performed, which by MPEP 2106.05(h) cannot integrate the abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, the additional elements, alone or in combination, do not provide significantly more than the abstract idea itself, because the particular field of use or particular technological environment (MPEP 2106.05(h)), implementation on a computer (MPEP 2106.05(f)), and the activity of data gathering (MPEP 2106.05(g)) cannot provide significantly more, as storing and retrieving information in memory is well understood, routine, and conventional (MPEP 2106.05(d)(II)(iv)) and the combination of additional elements does not provide an inventive concept. Thus, the claim is ineligible.
Claim 2, dependent upon Claim 1, further recites no additional abstract ideas. However:
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
the neural network has a softmax output in a middle layer, which merely specifies the particular field of use or particular technological environment in which the abstract idea is to be performed, which by MPEP 2106.05(h) cannot integrate the abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because the particular field of use or particular technological environment (MPEP 2106.05(h)) cannot provide significantly more. Thus, the claim is ineligible.
Claim 3, dependent upon Claim 1, further recites
generate a first feature vector from the target information in the training data, which is an evaluation or judgement that can be performed in the human mind;
generate the condition vector as the condition from the first feature vector, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
The system further comprises a processor that is configured to execute instructions, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because implementation on a computer (MPEP 2106.05(f)) cannot provide significantly more. Thus, the claim is ineligible.
Claim 4, dependent upon Claim 3, further recites
generate a second feature vector as the feature quantity from the input information in the training data, which is an evaluation or judgement that can be performed in the human mind;
connecting the condition vector to the second feature vector, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim 7, dependent upon Claim 1, further recites no additional abstract ideas. However:
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
the neural network includes an encoder and a decoder, which merely specifies the particular field of use or particular technological environment in which the abstract idea is to be performed, which by MPEP 2106.05(h) cannot integrate the abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because the particular field of use or particular technological environment (MPEP 2106.05(h)) cannot provide significantly more. Thus, the claim is ineligible.
Claim 8 recites A method, thus a process, one of the four statutory categories of patentable subject matter. However, Claim 8 recites precisely the abstract ideas and additional elements of Claim 1. Therefore, Step 2A Prong 1, Step 2A Prong 2, and Step 2B analyses remain the same, thus Claim 8 is rejected as subject-matter ineligible for reasons set forth in the rejections of Claim 1.
Claim 9 recites A non-transitory recording medium, thus an article of manufacture, one of the four statutory categories of patentable subject matter. However, Claim 9 recites on which an information learning program that allows a computer to execute an information learning method is recorded, the information learning method comprising precisely the abstract ideas and additional elements of Claim 1. Therefore, Step 2A Prong 1 analysis remains the same. As for Step 2A Prong 2 and Step 2B: performance on a computer cannot integrate an abstract idea into a practical application (Step 2A Prong 2) nor provide significantly more than the abstract idea itself (Step 2B) (MPEP 2106.05(f)), and thus Claim 9 is rejected as subject-matter ineligible for reasons set forth in the rejections of Claim 1.
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.
Claims 1-4 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Wang et al., "KDGAN: Knowledge Distillation with Generative Adversarial Networks" (hereinafter Wang).
Regarding Claim 1, Wang teaches:
A system for probabilistic sampling temperature of activation in conditional generative neural networks, the system comprising: at least one memory that is configured to store information and instructions; and at least one processor that is configured to execute the instructions, the instructions causing the at least one processor to: (Wang, p.6 last line, 3 The code and the data are made available at https://github.com/xiaojiew1/KDGAN/.) (Note: code requires memory and a processor to execute)
generate a condition vector from training data that are inputted to a neural network; (Wang, Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
) (Note: st and sc each correspond to the generated condition vector; the KDGAN architecture corresponds to a neural network)
connect the condition vector to a feature quantity of the training data, wherein the connecting the condition vector comprises: (Wang, Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
) (Note: pQt (y|x) and x correspond to training data; the categorical distribution pQt (y|x) corresponds to a feature quantity of the training data (for the classifier), thus connecting the condition to a feature quantity of the training data)
obtaining, from target information in the training data, a feature vector and apply a softmax activation with a temperature parameter T to generate the condition vector, the condition vector being a one-hot vector;
(Wang, p.6 ¶2,
PNG
media_image2.png
236
821
media_image2.png
Greyscale
PNG
media_image3.png
32
799
media_image3.png
Greyscale
Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
[p.4 last ¶] We formulate KDGAN as a minimax game with a classifier C, a teacher T, and a discriminator D. Similar to the classifier C, the teacher T generates pseudo labels based on a categorical distribution pQt (y|x) = softmax(f(x, y)) where f(x, y) is also a scoring function.) (Note: the KDGAN architecture corresponds to a neural network; pQt is the equation for st (the condition vector); a categorical distribution corresponds to a one-hot vector; y corresponds to the target information)
during learning, sampling the temperature parameter T for the softmax from a uniform distribution U(Tmin, Tmax) by a temperature sampling unit,…
(Wang, p.6 ¶2,
PNG
media_image2.png
236
821
media_image2.png
Greyscale
PNG
media_image3.png
32
799
media_image3.png
Greyscale
) (Note: the Gumbel distribution g is sampled by drawing u ~Uniform and thus corresponds to the uniform distribution)
…independently for each training step;
(Wang [fig.2]
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
)(Note: each yc calculated in each training loop from classifier to discriminator corresponds to each training step)
connecting the condition vector to the feature quantity of an input information by concatenation in a channel direction to form a combined feature;
(Wang, Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
) (Note: pQt (y|x) corresponds to training data; x corresponds to the input information; the categorical distribution pQt (y|x) corresponds to a feature quantity of the training data (for the classifier); the arrows from x connect (and concatenate) the condition to a feature quantity of the training data; yc is generated from corresponds to the combined feature; the arrow from classifier to discriminator which generates yc corresponds to the channel direction)
optimize a parameter of the neural network by using the feature quantity to which the condition vector is connected; and compute an error function between an output decoded from the combined feature and the target information and update the neural network parameters by back-propagation using the gradient of the error function. (Wang, Algorithm 1 line 12,
[p.6 ¶6] In addition to improving the training of C, we also apply the same techniques to improve the training of T. We update D with the back-propagation algorithm [37] (detailed in Appendix B). The overall logic of the KDGAN training is summarized in Algorithm 1. The three players can be first pretrained separately and then trained alternatively via minibatch stochastic gradient descent) (Note: updating D corresponds to updating the neural network parameters; the loss function in line 6 corresponds to the error function between the output (yc) and the target information (y))
Regarding Claim 2, Wang respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Wang further teaches: The system according to claim 1, wherein the neural network has a softmax output in a middle layer. (Wang, Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
[p.4 last ¶] We formulate KDGAN as a minimax game with a classifier C, a teacher T, and a discriminator D. Similar to the classifier C, the teacher T generates pseudo labels based on a categorical distribution pQt (y|x) = softmax(f(x, y)) where f(x, y) is also a scoring function.) (Note: the KDGAN architecture corresponds to a neural network)
Regarding Claim 3, Wang respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Wang further teaches:
The system according to claim 1, wherein the instructions further cause the at least one processor to generate a first feature vector from a target information in the training data, and generate the condition vector as the condition from the first feature vector. (Wang, Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
) (Note: pQt (y|x) corresponds to a first feature vector; y corresponds to a target information; yc is generated from the inputted pQt(y|x) and thus corresponds to the condition vector)
Regarding Claim 4, Wang respectively teaches and incorporates the claimed limitations and rejections of Claim 3. Wang further teaches:
The system according to claim 3, wherein the instructions further cause the at least one processor to generate a second feature vector as the feature quantity from an input information in the training data, and connecting the condition vector to the second feature vector. (Wang, Fig.2,
PNG
media_image1.png
388
1083
media_image1.png
Greyscale
) (Note: pc (y|x) corresponds to a second feature vector; x corresponds to input information; pc(y|x) is connected to pQt and thus corresponds to connected the condition vector to the second feature vector)
Claim Rejections - 35 USC § 103
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 7 is rejected under 35 U.S.C. 103 as being unpatentable over Wang in view of Wang_Supp et al., "KDGAN: Knowledge Distillation with Generative Adversarial Networks, Supplemental" (hereinafter Wang_Supp).
Regarding Claim 7, Wang respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Wang does not teach, but Wang_Supp further teaches:
The system according to claim 1, wherein the neural network includes an encoder (Wang_Supp [p.15, C. Network Architectures] We implement the scoring function s(x, y) as a LeNet [27]. The architecture of the LeNet is given by
1. An input layer of a 28×28 grayscale image.
2. A convolutional layer with 32 kernels of size 5×5 and stride 1.
3. A max pooling layer with size 2×2 and stride 2.
4. A convolutional layer with 64 kernels of size5×5 and stride 1.
5. A max pooling layer with size 2×2 and stride 2.)
and a decoder. (Wang_Supp []
6. A fully connected layer with 1024 neurons.
7. A softmax layer with 10 classes.)
Wang and Wang_supp are analogous to the present invention because both are from the same field of endeavor of making a conditional inference. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the LeNet containing an encoder and decoder from Wang_supp to Wang’s categorical distribution network. The motivation would be “to conduct experiments in deep model compression and image tag recommendation tasks” (Wang_supp, p.15 section C).
Independent Claim 8 recites a method to perform precisely the methods of Claim 1. Thus, Claim 8 is rejected for reasons set forth in Claim 1.
Independent Claim 9 recites A non-transitory recording medium on which an information learning program that allows a computer to execute an information learning method is recorded, the information learning method comprising (Wang, p.6 last line, 3 The code and the data are made available at https://github.com/xiaojiew1/KDGAN/.) (Note: code requires memory and a processor to execute) to perform precisely the methods of Claim 1. Thus, Claim 9 is rejected for reasons set forth in Claim 1.
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 JOSEP HAN whose telephone number is (703)756-1346. The examiner can normally be reached Mon-Fri 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, Kakali Chaki can be reached on (571) 272-3719. 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.
/J.H./Examiner, Art Unit 2122
/KAKALI CHAKI/Supervisory Patent Examiner, Art Unit 2122
Prosecution Insights