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 .
Claims 1-20 are presented for examination.
Continued Examination under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on October 22, 2025 has been entered.
Response to Amendment
Applicant’s amendment has obviated most, but not all, of the specification objections. To the extent that an objection or rejection appears in the previous Office Action(s) but not this Office Action, that objection or rejection is withdrawn. To the extent that it appears both in a previous Office Action(s) and this Office Action, the objection or rejection is maintained.
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Specification
Examiner objects to the specification for containing the following informalities:
In paragraph 3, “in which they are implement” should be “in which they are implemented”.
In paragraph 101, “represents accumulated” should be “represents an accumulated”.
In paragraph 189, the trademarked term BLU-RAY should be given proper attribution as such.
Appropriate correction is required.
Claim Rejections - 35 USC § 101
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to and abstract idea without significantly more. The analysis of the claims will follow the 2019 Revised Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (“2019 PEG”).
Claim 1
Step 1: The claim recites a method; therefore, it is directed to the statutory category of process.
Step 2A Prong 1: The claim recites, inter alia:
[C]alculating individual update values for a weight assigned to a connection relationship between nodes included in a neural network, where the individual update values are below a threshold precision of the weight: This limitation could encompass the mental calculation of the update values below a threshold for a weight. Additionally or alternatively, this limitation represents a mathematical calculation.
[G]enerating an accumulated update value for updating the weight … when the accumulated update value is equal to or greater than a threshold value by adding the individual update values: This limitation recites a mathematical concept of adding update values that could be performed in the mind given a sufficiently small dataset.
[T]raining the neural network by updating the weight using the accumulated update value in response to the accumulated update value being equal to or greater than the threshold value, wherein the threshold value is a value of 2n of an n-th bit of the weight, where the n-th bit is a bit of lesser significance than a bit in the weight representing a largest magnitude bit among all bits of the weight: The specification, at paragraph 68, indicates that the training of the network may be performed through backpropagation, which is a mathematical concept. As such, this limitation recites a mathematical concept of performing a comparison between a threshold and an accumulated update value that is greater than the maximum decimal number representable with n bits and updating the weights through training based thereon.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The claim further recites that the method is a “processor-implemented neural network” method; that the update values are below a threshold precision “during training of the neural network”; and that the weight update occurs “during the training”. However, these amount to mere instructions to apply the judicial exception using a generic computer. MPEP § 2106.05(f).
Step 2B: The claim does not contain significantly more than the judicial exception. The analysis at this step mirrors that of step 2A, prong 2. As an ordered whole, the claim recites a mathematical algorithm for training a neural network. Nothing in the claim provides significantly more than this. As such, the claim is not patent eligible.
Claim 2
Step 1: A process, as above.
Step 2A Prong 1: The claim recites that “the updating comprises updating the weight by adding an effective update value to the weight.” This limitation recites a mathematical concept of adding an update value to a weight.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 1 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 1 analysis.
Claim 3
Step 1: A process, as above.
Step 2A Prong 1: The claim recites:
[G]enerating a new accumulated update value by subtracting, corresponding to the updating of the weight, the effective update value from the accumulated update value: This limitation recites a mathematical concept of subtracting an effective update value from an accumulated update value.
[A]dding another individual update value to the new accumulated update value: This limitation recites a mathematical concept of adding an individual update value to a new accumulated update value.
[U]pdating the updated weight using the new accumulated update value in response to the new accumulated update value being equal to or greater than the threshold value: This limitation recites the mathematical concept of comparing an accumulated update value with a threshold value and updating the weight if the threshold is reached.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 2 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 2 analysis.
Claim 4
Step 1: A process, as above.
Step 2A Prong 1: The claim recites “adjusting a size of the accumulated update value, by a factor, based on a comparison between a second threshold value and either one or both of an average value of the individual update values and the accumulated update value, wherein the updating comprises updating the weight using the adjusted accumulated update value”. This limitation recites the mathematical concept of comparing a threshold value to an average of individual update values and/or an accumulated update value, adjusting a size of an accumulated update value based on this comparison, and using the result of the adjustment to update the weight. This procedure could also be performed mentally with sufficiently simple numbers.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 1 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 1 analysis.
Claim 5
Step 1: A process, as above.
Step 2A Prong 1: The claim recites that “the n-th bit is a least significant effective bit of the weight”. The weight update remains a mathematical concept under this further assumption.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 1 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 1 analysis.
Claim 6
Step 1: A process, as above.
Step 2A Prong 1: The claim recites that “when the weight is a fixed point value, the least significant effective bit of the weight is the least significant bit of the weight, and when the weight is a floating point value, the least significant effective bit of the weight is based on the least significant bit of the weight and a bias of the weight”. Updating the weight remains a mathematical concept under these further assumptions.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 5 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 5 analysis.
Claim 7
Step 1: A process, as above.
Step 2A Prong 1: The claim recites “determining whether the accumulated update value is equal to or greater than the threshold value at a predetermined update period”. This limitation recites the mathematical concept of comparing an accumulated update value to a threshold value; this limitation could also be performed in the mind.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 1 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 1 analysis.
Claim 8
Step 1: A process, as above.
Step 2A Prong 1: The claim recites “accumulating the individual update values until a next update period in response to a result of the determining being that the accumulated update value is smaller than the threshold value”. This limitation is directed to the mathematical concept of comparing an accumulated update value to a threshold value and continuing to accumulate (i.e., add) update values until the accumulated value reaches the threshold. This limitation could also be performed mentally.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 7 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 7 analysis.
Claim 9
Step 1: The claim recites a method; therefore, it is directed to the statutory category of processes.
Step 2A Prong 1: The claim recites, inter alia:
[C]alculating individual update values for a weight assigned to a connection relationship between nodes included in a neural network, where the individual update values are below a threshold precision of the weight: This limitation recites a mathematical concept of calculating update values below a threshold precision for a weight connecting nodes in a neural network.
[G]enerating an accumulated update value for updating the weight … when the accumulated update value is equal to or greater than a threshold value by accumulating the individual update values: This limitation recites the mathematical concept of accumulating (i.e., adding) update values; this could also be performed mentally given a sufficiently small dataset.
[T]training the neural network by updating the weight using the accumulated update value in response to the accumulated update value being equal to or greater than the threshold value, wherein: the weight is a floating point value comprising a first sign bit, a first exponent part, a first mantissa part, and a first bias; [and] the accumulated update value is a floating point value comprising a second sign bit, a second exponent part, a second mantissa part, and a second bias: As noted above, paragraph 68 of the specification indicates that the training may be performed via backpropagation. As such, this limitation recites the mathematical concept of comparing an accumulated update value with the claimed characteristics to a threshold value and using the accumulated update value to update a weight with the claimed characteristics as part of training a neural network.
[A]dding an effective value of the accumulated update value included in an effective number range of the weight to the weight: This limitation recites a mathematical concept of adding an effective value to a weight; this could also be performed mentally with sufficiently small numbers.
[A]djusting the second bias of the accumulated update value: This limitation could encompass mentally adjusting the second bias; to the extent that “adjusting” entails addition or subtraction, this could also recite a mathematical concept.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The claim further recites that the method is a “processor-implemented neural network” method; that the accumulation of the individual update values occurs “in an accumulation buffer”; that the update values are below a threshold precision “during training of the neural network”; and that the weight update occurs “during the training”. . However, these are mere instructions to apply the judicial exception using a generic computer. MPEP § 2106.05(f).
Step 2B: The claim does not contain significantly more than the judicial exception. The analysis at this step mirrors that of step 2A, prong 2. As an ordered whole, the claim is directed to a mathematical concept of training a neural network using floating point arithmetic. Nothing in the claim provides significantly more than this. As such, the claim is not patent eligible.
Claim 10
Step 1: A process, as above.
Step 2A Prong 1: The claim recites “increasing the second bias in response to the second exponent of the accumulated update value being greater than the threshold value; and decreasing the second bias in response to the accumulated update value being smaller than a second threshold value.” These limitations recite the mathematical concept of comparing an accumulated update value to a threshold value and increasing or decreasing the second bias based on the comparison. This limitation could also be performed mentally.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 9 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 9 analysis.
Claim 11
Step 1: A process, as above.
Step 2A Prong 1: The claim recites that “the second threshold value is 1/b times the threshold value; and b is a natural number.” This limitation recites the mathematical concept of multiplying a threshold value by a constant, which could be performed mentally.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 10 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 10 analysis.
Claim 12
Step 1: A process, as above.
Step 2A Prong 1: The claim recites that “a least significant effective bit of the weight is based on a least significant bit of the weight and the first bias”. Updating the weights remains a mathematical concept under these further assumptions.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See claim 9 analysis.
Step 2B: The claim does not contain significantly more than the judicial exception. See claim 9 analysis.
Claims 13-20
Step 1: The claims recite an apparatus comprising one or more processors; therefore, they are directed to the statutory category of machines.
Step 2A Prong 1: The claims recite the same abstract ideas as in claims 1-8, respectively.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The analysis at this step is the same as in claims 1-8, respectively, except insofar as these claims recite a “neural network apparatus, the apparatus comprising: one or more processors configured to [perform the method]”. However, this is a mere instruction to apply the judicial exception using a generic computer. MPEP § 2106.05(f).
Step 2B: The claim does not contain significantly more than the judicial exception. The analysis at this step is the same as in claims 1-8, respectively, except insofar as these claims recite a “neural network apparatus, the apparatus comprising: one or more processors configured to [perform the method]”. However, this is a mere instruction to apply the judicial exception using a generic computer. MPEP § 2106.05(f).
Response to Arguments
Applicant's arguments filed October 22, 2025 (“Remarks”) have been fully considered but they are, except insofar as rendered moot by the withdrawal of a ground of rejection, not persuasive.
Applicant first argues that the claims as amended are ineligible because (a) Examiner is allegedly improperly requiring that the benefit of the claimed subject matter be explicitly recited in the claims; and (b) the claims as amended are now allegedly directed to an improvement in the technical field of using low-precision updates in neural networks. Remarks at 7-11. However, Examiner never asserted that such benefit must be recited in the claims, but rather that the benefit must be reflected in the claim language. “After the examiner has consulted the specification and determined that the disclosed invention improves technology, the claim must be evaluated to ensure the claim itself reflects the disclosed improvement in technology.” MPEP § 2106.05(a), fourth paragraph. Here, the benefit cannot be reflected in the claim language because the added language providing that purported benefit is part of the judicial exception itself, and “the judicial exception alone cannot provide the improvement.” Id. at sixth paragraph. That is, to the extent that the claim purports to be an improvement to neural network training, the claim merely improves the judicial exception itself because the training is recited in such a way that it may be considered a mathematical concept. Compare Example 47, claim 2.
The remarks with respect to the art rejections, Remarks at 11-17, are moot in light of the withdrawal of that ground of rejection.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RYAN C VAUGHN whose telephone number is (571)272-4849. The examiner can normally be reached M-R 7:00a-5:00p ET.
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, Kamran Afshar, can be reached at 571-272-7796. 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.
/RYAN C VAUGHN/ Primary Examiner, Art Unit 2125