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
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 03/04/2026 has been entered.
Response to Amendments
Claims 1-15 remain pending in the application.
Claims 1, 4, 6, 9, 11, and 14 have been amended.
The amendment filed 03/04/2026 is sufficient to overcome the 102 rejections of claims 1-3, 6-8, and 11-13 as being anticipated by Hersam and the 103 rejections of claims 4-5, 9-10, and 14-15 over Hersam in view of Creed. The previous rejections have been withdrawn.
Response to Arguments
Argument 1, regarding the prior art rejections, applicant argues that Hersam does not teach analyzing/modifying the ANN model based on a target-architecture constraint rule set and a performance metric as claim 1 now recites. Applicant argues that the cited portions of Hersam are directed towards updating weights and training of an artificial neural network and not directed towards analyzing/modifying the ANN based on a target-hardware architecture constraint rule set and performance metric.
Examiner notes that Pfeil et al (Pub. No.: US 20220327354 A1), hereafter Pfeil, teaches a set of rules specifying ANN model constraints of the target hardware architecture (an artificial neural network is executed in parts due to hardware and energy constraints. This includes a limited number of iterations an artificial neural network may execute which is defined by hardware constraints and limited processing time, P0002, P0008, P0009, P0039)… generating an output including one or more modifications for the ANN model based on the set of rules (parameters of the ANN may be optimized in view of a cost function that outlines energy limits, P0024, P0038, P0039).
Hersam teaches receiving an artificial neural network (ANN) model to run on a target hardware architecture (Hardware implements ANN, P0079); analyzing the ANN model based on a set of rules…and a performance metric for the ANN model (Learning rules are associated with learning behavior during hardware operation, and these learning rules are used to analyze ANNs. Learning rules are associated with hardware operation and hardware implementation of the ANN, P0172. In view of P0027 of the specification of the instant application, a performance metric may be power consumption of a model. Energy consumption is a metric used to evaluate performance of the ANN, P0173); and generating an output including one or more modifications for the ANN model based on the set of rules and the performance metric for the ANN model (Weights of the ANN are updated according to the learning rules, P0172. Weights of the ANN may be updated based on the energy consumed per switching cycle for the purpose of reducing power consumption, P0173).
Thus, the combination of Hersam in view of Pfeil teaches the limitations of amended claims 1, 6, and 11. The full prior art rejections are outlined below.
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.
Claims 1-3, 6-8, and 11-13 are rejected under 35 U.S.C. 103 over Hersam et al (Pub. No.: US 20210098611 A1), hereafter Hersam in view of Pfeil et al (Pub. No.: US 20220327354 A1), hereafter Pfeil.
Regarding claims 1, 6, and 11, Hersam teaches receiving an artificial neural network (ANN) model to run on a target hardware architecture (Hardware implements ANN, P0079); analyzing the ANN model based on a set of rules…and a performance metric for the ANN model (Learning rules are associated with learning behavior during hardware operation, and these learning rules are used to analyze ANNs. Learning rules are associated with hardware operation and hardware implementation of the ANN, P0172. In view of P0027 of the specification of the instant application, a performance metric may be power consumption of a model. Energy consumption is a metric used to evaluate performance of the ANN, P0173); and generating an output including one or more modifications for the ANN model based on the set of rules and the performance metric for the ANN model (Weights of the ANN are updated according to the learning rules, P0172. Weights of the ANN may be updated based on the energy consumed per switching cycle for the purpose of reducing power consumption, P0173).
Hersam does not appear to explicitly teach “specifying ANN model constraints of the target hardware architecture”.
Pfeil teaches specifying ANN model constraints of the target hardware architecture (an artificial neural network is executed in parts due to hardware and energy constraints. This includes a limited number of iterations an artificial neural network may execute which is defined by hardware constraints and limited processing time, P0002, P0008, P0009, P0039)… generating an output including one or more modifications for the ANN model based on the set of rules and the performance metric for the ANN model (parameters of the ANN may be optimized in view of a cost function that outlines energy limits, P0024, P0038, P0039).
Accordingly, it would have been obvious to a person having ordinary skill in the
art before the effective filing date of the claimed invention, having the teachings of
Hersam and Pfeil before them, to include Pfeil’s specific teaching of adjusting an artificial neural network’s parameters based on energy limits imposed by hardware
in Hersam’s method of semiconductor fabrication. One would have been motivated to make such a combination of adjusting an artificial neural network’s parameters based on energy limits imposed by hardware (see Pfeil P0038-P0039), and updating weights of the ANN based on the energy consumed per switching cycle (see Hersam P0042) to improve energy efficiency of hardware (see Pfeil P0018).
Regarding claims 2, 7, and 12, Hersam in view of Pfeil teaches the limitations of claims 1, 6, and 11 as outlined above. Hersam further teaches further comprising implementing at least one modification of the one or more modifications in the ANN model to generate an updated model (Weights of ANN model are updated, P0172).
Regarding claims 3, 8, and 13, Hersam in view of Pfeil teaches the limitations of claims 2, 7, and 12 as outlined above. Hersam further teaches further comprising operating the updated model on the target hardware architecture to generate an inference (A classification (inference) may be performed based on the hardware implementation of the ANN, P0005, P0079).
Claims 4-5, 9-10, and 14-15 are rejected under 35 U.S.C. 103 as being unpatentable over Hersam in view of Pfeil and further in view of Creed et al (Pub. No.: US 20210081717 A1), hereafter Creed.
Regarding claims 4, 9, and 14, Hersam in view of Pfeil teaches the limitations of claims 1, 6, and 10 as outlined below. Hersam does not appear to explicitly teach parsing a representation for the ANN model; determining a set of nodes corresponding to the representation; and applying a set of constraints for each node of the set of nodes based on the set of rules.
Creed teaches parsing a representation for the ANN model; determining a set of nodes corresponding to the representation (Nodes may be used to represent an ANN, P0084-P0085); and applying a set of constraints for each node of the set of nodes based on the set of rules, the set of constraints specified in the set of rules (Rules may be applied to the nodes that dictate the relationship between nodes, P0085-P0086).
Accordingly, it would have been obvious to a person having ordinary skill in the
art before the effective filing date of the claimed invention, having the teachings of
Hersam, Pfeil, and Creed before them, to include Creed’s specific teachings of nodes being used to represent an ANN and applying rules to nodes to dictate their relationship in Hersam’s method of semiconductor fabrication. One would have been motivated to make such a combination of nodes being used to represent an ANN implemented within hardware and applying rules to nodes to dictate their relationship (see Creed P0084-P0086, P0053) and evaluating hardware operation with a set of learning rules to update weights of an ANN (see Hersam P0172) to improve predictions made by the ANN (see Creed P0082).
Regarding claims 5, 10, and 15, Hersam in view of Pfeil and further in view of Creed teaches the limitations of claims 1, 6, and 10 as outlined below. Hersam does not appear to explicitly teach receiving a graph representation of the ANN model; identifying one or more patterns within the graph representation; applying the set of rules to the one or more patterns to determine nodes associated with the one or more patterns; and outputting a modification for the nodes associated with the one or more patterns.
Creed teaches receiving a graph representation of the ANN model (A graph structure may be created for an ANN, P0084-P0085); identifying one or more patterns within the graph representation (Trends and patterns may be identified in the graph, P0172); applying the set of rules to the one or more patterns to determine nodes associated with the one or more patterns (Edges, or a pair of nodes with a defined relationship in the graph, associated with the identified trends and patterns are identified, P0172); and outputting a modification for the nodes associated with the one or more patterns (Edges that are associated with the trends and patterns have their training weights adjusted, P0172).
Accordingly, it would have been obvious to a person having ordinary skill in the
art before the effective filing date of the claimed invention, having the teachings of
Hersam and Creed before them, to include Creed’s specific teachings of representing an ANN by a graph and identifying nodes associated with patterns in the graph to be modified in Creed’s method of semiconductor fabrication. One would have been motivated to make such a combination of representing an ANN implemented within hardware by a graph and identifying nodes associated with patterns in the graph to be modified (see Creed P0084-P0085, P0172, P0053) and evaluating hardware operation with a set of learning rules to update weights of an ANN (see Hersam P0172).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ISHAN MOUNDI whose telephone number is (703)756-1547. The examiner can normally be reached 8:30 A.M. - 5 P.M..
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, Matthew Ell can be reached at (571) 270-3264. 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.
/I.M./Examiner, Art Unit 2141
/DANIEL T PELLETT/Primary Examiner, Art Unit 2121
Prosecution Insights