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 . under the first inventor to file provisions of the AIA .
Claims 1-20 are pending and are being examined.
Examiner's Note
The Examiner respectfully requests of the Applicant in preparing responses, to fully consider the entirety of the reference(s) as potentially teaching all or part of the claimed invention. It is noted, REFERENCES ARE RELEVANT AS PRIOR ART FOR ALL THEY CONTAIN. “The use of patents as references is not limited to what the patentees describe as their own inventions or to the problems with which they are concerned. They are part of the literature of the art, relevant for all they contain.” In re Heck, 699 F.2d 1331, 1332-33, 216 USPQ 1038, 1039 (Fed. Cir. 1983) (quoting In re Lemelson, 397 F.2d 1006, 1009, 158 USPQ 275, 277 (CCPA 1968)). A reference may be relied upon for all that it would have reasonably suggested to one having ordinary skill in the art, including non-preferred embodiments (see MPEP 2123). The Examiner has cited particular locations in the reference(s) as applied to the claim(s) above for the convenience of the Applicant. Although the specified citations are representative of the teachings of the art and are applied to the specific limitations within the individual claim(s), typically other passages and figures will apply as well.
Claim Rejections - 35 USC § 101
101 Rejection
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 1-20 is rejected under 35 USC § 101 because the claimed invention is directed to non-statutory subject matter
Step 1 Analysis:
Claims 1-10 are directed to a method which is directed to a process, one of the statutory categories. Claims 11-15 are directed to a computing device, which is directed to a machine, one of the statutory categories. Claims 16-20 are directed to a non-transitory computer storage medium, which is directed to a product, one of the statutory categories.
Regarding Claim 1:
Claim 1 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 2A Prong 1 Analysis:
Claim 1 recites in part process steps which, under the broadest reasonable interpretation, are a series of mental processes including an observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper. If a claim, under its broadest reasonable interpretation, covers a mental process or a mathematical concept but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas. The claim recites in part:
identifying a type of accelerators of multiplication and accumulation operations Under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator identifying a type of accelerator). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
adjusting a weight matrix of an artificial neural network based on energy consumption characteristics of the type of accelerators Under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator modifying the weights of a model). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
Configuring ….. the weight matrix having been adjusted according to the energy consumption characteristics Under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator configuring a weight matrix in a device). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
Step 2A Prong 2 Analysis:
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
….. in a computing device having an accelerator of the type ….. is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
Step 2B Analysis:
Claim 1 does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above, the additional elements of:
….. in a computing device having an accelerator of the type ….. is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
For the reasons above, claim 1 is rejected as being directed to non-patentable subject matter under §101.
The additional limitations of the dependent claims contain no additional elements that provide a practical application or amount to significantly more than the abstract idea and are addressed briefly below.
For the reasons above, claim 1 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 2:
Step 2A Prong 1:
Accelerating ….. multiplication and accumulation operations under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator accelerating the execution of operations). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
….. using the accelerator of the type ….. is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
in computations of the artificial neural network performed using the weight matrix configured in the computing device is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
wherein the adjusting of the weight matrix includes training of the weight matrix according to a training dataset is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
….. using the accelerator of the type ….. is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
in computations of the artificial neural network performed using the weight matrix configured in the computing device is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
wherein the adjusting of the weight matrix includes training of the weight matrix according to a training dataset is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 2 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 3:
Step 2A Prong 1:
wherein the training of the weight matrix includes reducing a loss function under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator decreasing the difference between an actual number and a predicted number). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
associated with the energy consumption characteristics is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
associated with the energy consumption characteristics is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 3 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 4:
Step 2A Prong 1: The claim recites the same mental processes as the independent claim by dependency.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein accelerators of the type are implemented using microring resonators as computing elements for multiplication is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
the loss function is configured to penalize small weights more than large weights is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein accelerators of the type are implemented using microring resonators as computing elements for multiplication is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
the loss function is configured to penalize small weights more than large weights is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 4 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 5:
Step 2A Prong 1: The claim recites the same mental processes as the independent claim by dependency.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein accelerators of the type are implemented using memristors as computing elements for multiplication is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
the loss function is configured to penalize large weights more than small weights is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein accelerators of the type are implemented using memristors as computing elements for multiplication is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
the loss function is configured to penalize large weights more than small weights is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 5 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 6:
Step 2A Prong 1: The claim recites the same mental processes as the independent claim by dependency.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein accelerators of the type are implemented using synapse memory cells as computing elements for multiplication is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
the loss function is configured to penalize a first type of bits more than a second type of bits in weights is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein accelerators of the type are implemented using synapse memory cells as computing elements for multiplication is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
the loss function is configured to penalize a first type of bits more than a second type of bits in weights is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 6 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 7:
Step 2A Prong 1: The claim recites the same mental processes as the independent claim by dependency.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein bits of the first type have a value of one; and bits of the second type have a value of zero is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein bits of the first type have a value of one; and bits of the second type have a value of zero is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 7 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 8:
Step 2A Prong 1: The claim recites the same mental processes as the independent claim by dependency.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein the adjusting of the weight matrix includes retraining an input weight matrix according to a pruning selection to suppress a pattern of weights in the input weight matrix is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein the adjusting of the weight matrix includes retraining an input weight matrix according to a pruning selection to suppress a pattern of weights in the input weight matrix is recited at a high level of generality and amounts to no more than generally linking the use of a judicial exception to a particular technological environment or field of use MPEP 2106.05(h). As explained by the Supreme Court, a claim directed to a judicial exception cannot be made eligible "simply by having the applicant acquiesce to limiting the reach of the patent for the formula to a particular technological use." Diamond v. Diehr, 450 U.S. 175, 192 n.14, 209 USPQ 1, 10 n. 14 (1981). Thus, limitations that amount to merely indicating a field of use or technological environment in which to apply a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
For the reasons above, claim 8 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 9:
Step 2A Prong 1:
modifying a first portion of the input weight matrix and adjusting a second portion of the input weight matrix under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator adjusting portions of a weight matrix). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
reduce differences between outputs generated using the input weight matrix and outputs generated using a re-trained weight matrix under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator decreasing the difference between two predicted numbers). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein the re-training includes is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f))..
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein the re-training includes is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
For the reasons above, claim 9 is rejected as being directed to non-patentable subject matter under §101.
Dependent claim 10:
Step 2A Prong 1:
determining an accuracy performance level of the re-trained weight matrix under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator determining how accurate is a predicted value from a retrained model compared to an expected value). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
determining an energy performance level of the re-trained weight matrix under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator determining how much energy a retrained model consumes). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas
evaluating a combined performed level based on the accuracy performance level and the energy performance level under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator evaluating the cost benefit to retrain a model and how much more accurate it could become). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
Searching for a weight selection and modification solution to improve or optimize the combined performed level under the broadest reasonable interpretation, this limitation is a process step that covers a mental process including observation, evaluation, judgment or opinion that could be performed in the human mind or with the aid of pencil and paper (such as an operator searching for the best level of retraining to produce the most cost effective accurate model). If a claim, under its broadest reasonable interpretation, covers a mental process but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas.
Step 2A Prong 2: The judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of:
wherein the re-training further includes is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f))..
Step 2B: In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
wherein the re-training further includes is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
For the reasons above, claim 10 is rejected as being directed to non-patentable subject matter under §101.
Claims 11-12 are the computing device claims corresponding to the method claims 1-2 respectively, therefore, they are rejected based upon the same rationale as the rejection of claims 1-2.
Claim 13 is the computing device claim corresponding to the method claim 4 respectively, therefore, it is rejected based upon the same rationale as the rejection of claim 4.
Claim 14 is the computing device claim corresponding to the method claim 5 respectively, therefore, it is rejected based upon the same rationale as the rejection of claim 5.
Claim 15 is the computing device claim corresponding to the method claim 6 respectively, therefore, it is rejected based upon the same rationale as the rejection of claim 6.
Claim 16 is the non-transitory computer storage medium claim corresponding to the method claims 1 and 9, therefore, it is rejected based upon the same rationale as the rejection of claims 1 and 9.
Claim 17 is the non-transitory computer storage medium claim corresponding to the method claims 2 and 10, therefore, it is rejected based upon the same rationale as the rejection of claims 2 and 10.
Claim 18 is the non-transitory computer storage medium claim corresponding to the method claim 4, therefore, it is rejected based upon the same rationale as the rejection of claim 4.
Claim 19 is the non-transitory computer storage medium claim corresponding to the method claim 5, therefore, it is rejected based upon the same rationale as the rejection of claim 5.
Claim 20 is the non-transitory computer storage medium claim corresponding to the method claim 6, therefore, it is rejected based upon the same rationale as the rejection of claim 6.
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 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 1-2, 8-9, 11-12, and 16 are rejected under 35 U.S.C. 102 as being unpatentable over YANG (Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruning).
Regarding claim 1, YANG teaches identifying a type of accelerators of multiplication and accumulation operations ([Page 6072-6073, Sec. 2.1] Multiply-and-accumulate (MAC) operations in CONV and FC layers account for over 99% of total operations in state-of-the-art CNNs [3, 17, 19, 24], and therefore dominate both processing runtime and energy consumption……. This is because modern hardware processors implement multiple levels of memory hierarchy, e.g., DRAM and multi-level buffers, to amortize the energy cost of memory accesses.)
adjusting a weight matrix of an artificial neural network based on energy consumption characteristics of the type of accelerators ([Page 6074, Se. 4.1] Thus, in order to maximize the overall energy reduction, we prune the layers that consume the most energy first. The examiner notes that YANG teaches energy-aware pruning of neural networks, a process that requires adjusting the weight matrix of the pruned layers to zero).
configuring, in a computing device having an accelerator of the type, the weight matrix having been adjusted according to the energy consumption characteristics ([Page 6072, Col. 2] for a compact CNN, such as GoogLeNet, the proposed pruning method can still reduce energy consumption by 1.6×. The pruned models will be released at [23]. The examiner notes that YANG deploys the model with the adjusted weight matrix to an accelerator.)
Regarding claim 2, YANG teaches accelerating, using the accelerator of the type, multiplication and accumulation operations in computations of the artificial neural network performed using the weight matrix configured in the computing device ([Page 6072, Energy-Aware pruning] As a result, for AlexNet, we can reduce energy consumption by 3.7x after pruning, which is 1.7× lower than pruning with the popular network pruning method proposed in [8]. The examiner notes that YANG teaches deploying a model with an adjusted weight matrix to an accelerator to achieve a reduction in energy consumption by 3.7x.)
wherein the adjusting of the weight matrix includes training of the weight matrix according to a training dataset ([Page 6072, Col. 1] Each layer is first pruned and the preserved weights are locally finetuned with a closed-form least-square solution to quickly restore the accuracy and increase the compression ratio.)
Regarding claim 8, YANG teaches wherein the adjusting of the weight matrix includes retraining an input weight matrix according to a pruning selection to suppress a pattern of weights in the input weight matrix ([Page 6075, Sec. 4.4] The previous two steps select a subset of weights to preserve, but do not change the values of the weights. In this step, we perform the least-square optimization on each filter to change the values of their weights to further reduce the output error and restore the network accuracy
PNG
media_image1.png
74
558
media_image1.png
Greyscale
where the subscript Si means choosing the non-pruned weights from the ith filter and the corresponding columns from Xi. The least-square problem has a closed-form solution, which can be efficiently solved. The examiner notes that YANG teaches re-training the model to suppress the weights of the pruned layers.)
Regarding claim 9, YANG teaches wherein the re-training includes modifying a first portion of the input weight matrix and adjusting a second portion of the input weight matrix to reduce differences between outputs generated using the input weight matrix and outputs generated using a re-trained weight matrix ([Page 6075, Sec. 4.4] The previous two steps select a subset of weights to preserve, but do not change the values of the weights. In this step, we perform the least-square optimization on each filter to change the values of their weights to further reduce the output error and restore the network accuracy
PNG
media_image1.png
74
558
media_image1.png
Greyscale
where the subscript Si means choosing the non-pruned weights from the ith filter and the corresponding columns from Xi. The least-square problem has a closed-form solution, which can be efficiently solved. The examiner notes that YANG teaches re-training the model to zero out the weights of the layers that were pruned and changing the values of the weights to further reduce the output error and restore the network accuracy.)
Claims 11-12 are the computing device claims corresponding to the method claims 1-2 respectively, therefore, they are rejected based upon the same rationale as the rejection of claims 1-2.
Claim 16 is the non-transitory computer storage medium claim corresponding to the method claims 1 and 9, therefore, it is rejected based upon the same rationale as the rejection of claims 1 and 9.
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 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.
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over YANG (Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruning), in view of VISWANATHAN (US20200175339A1).
Regarding claim 3, YANG teaches the method of claim 2, however, YANG is not relied upon to explicitly teach wherein the training of the weight matrix includes reducing a loss function associated with the energy consumption characteristics. On the other hand, VISWANATHAN teaches wherein the training of the weight matrix includes reducing a loss function associated with the energy consumption characteristics ([0047] In addition, joint optimization of classification/regression performance and power consumption (e.g., model training using a loss function consider accuracy and/or power consumption) can guarantee worst-case performance measures as relevant to maintaining available power for vehicle safety features. The examiner notes that YANG and VISWANATHAN are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s CNN to incorporate wherein the training of the weight matrix includes reducing a loss function associated with the energy consumption characteristics as taught by VISWANATHAN [0047] to guarantee worst-case performance measures as relevant to maintaining available power for vehicle safety features [0047]).
Claims 4, 13, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over YANG (Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruning), in view of SHIFLETT (US20240127040A1), further in view of MAJUMDAR (US20170061328A1).
Regarding claim 4, YANG teaches the method of claim 3, however, YANG is not relied upon to explicitly teach:
wherein accelerators of the type are implemented using microring resonators as computing elements for multiplication.
the loss function is configured to penalize small weights more than large weights.
On the other hand, SHIFLETT teaches wherein accelerators of the type are implemented using microring resonators as computing elements for multiplication ([0084] PIXEL is a mixed-signal photonic accelerator built using microring resonators for bitwise logical operations and Mach-Zehnder modulators for analog accumulation. The examiner notes that YANG and SHIFLETT are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s CNN implementation to incorporate wherein accelerators of the type are implemented using microring resonators as computing elements for multiplication as taught by SHIFLETT [0084] for bitwise logical operations and Mach-Zehnder modulators for analog accumulation [0084]).
Furthermore, MAJUMDAR teaches the loss function is configured to penalize small weights more than large weights ([0076] In some aspects, the cost function may include least absolute errors (LI-norm) or L0 norm regularization. The examiner notes that MAJUDMAR teaches a loss function that uses L1 regularization which drives smaller weights to zero faster than larger weights due to the absolute penalty added to each weight. The examiner notes that YANG and MAJUMDAR are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s loss function to incorporate the loss function is configured to penalize small weights more than large weights as taught by MAJUMDAR [0076] to penalizes low sparsity of the feature vector [0076]).
Claim 13 is the computer device claim corresponding to the method claim 4, therefore, it is rejected based upon the same rationale as the rejection of claim 4.
Claim 18 is the non-transitory computer storage medium claim corresponding to the method claim 4, therefore, it is rejected based upon the same rationale as the rejection of claim 4.
Claims 5-7, 14-15, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over YANG (Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruning), in view of QIN (Design of High Robustness BNN Inference Accelerator Based on Binary Memristors), further in view of XU (US20220164511A1).
Regarding claim 5, YANG teaches the method of claim 3, however, YANG is not relied upon to explicitly teach:
wherein accelerators of the type are implemented using memristors as computing elements for multiplication.
the loss function is configured to penalize large weights more than small weights.
On the other hand, QIN teaches wherein accelerators of the type are implemented using memristors as computing elements for multiplication ([Page 3435, Sec. I] The memristor can update and store the analog synaptic weight (represented by the device conductance) in situ. This nonvolatile feature may lead to a significant reduction in data movement latency, energy consumption, and the requirement in data bandwidth, providing a promising solution for off-chip inference acceleration. The examiner notes that YANG and QIN are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s CNN implementation to incorporate wherein accelerators of the type are implemented using memristors as computing elements for multiplication as taught by QIN [Page 3435, Sec. I] to reduce energy consumption [Page 3435, Sec. I]).
Furthermore, XU teaches the loss function is configured to penalize large weights more than small weights ([0110] For this new linear model, an L2 regularizer, such as a Ridge regularizer, is applied, with a weaker penalty strength compared to the A used in the MCP regression. The examiner notes that XU teaches a loss function that uses L2 regularization which penalizes larger weights more than smaller weights due to the squared penalty that is added to the weight. The examiner further notes that YANG and XU are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s loss function to incorporate the loss function is configured to penalize large weights more than small weights as taught by XU [0110] to further boost the model accuracy [0110]).
Regarding claim 6, YANG teaches the method of claim 3, however, YANG is not relied upon to explicitly teach:
wherein accelerators of the type are implemented using synapse memory cells as computing elements for multiplication.
the loss function is configured to penalize a first type of bits more than a second type of bits in weights.
On the other hand, QIN teaches wherein accelerators of the type are implemented using synapse memory cells as computing elements for multiplication ([Page 3435, Sec. I] The memristor can update and store the analog synaptic weight (represented by the device conductance) in situ. This nonvolatile feature may lead to a significant reduction in data movement latency, energy consumption, and the requirement in data bandwidth, providing a promising solution for off-chip inference acceleration. The examiner notes that YANG and QIN are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s CNN implementation to incorporate wherein accelerators of the type are implemented using synapse memory cells as computing elements for multiplication as taught by QIN [Page 3435, Sec. I] to reduce energy consumption [Page 3435, Sec. I]).
Furthermore, XU teaches the loss function is configured to penalize a first type of bits more than a second type of bits in weights ([0110] For this new linear model, an L2 regularizer, such as a Ridge regularizer, is applied, with a weaker penalty strength compared to the A used in the MCP regression. The examiner notes that XU teaches a loss function that uses L2 regularization which penalizes larger weights more than smaller weights due to the squared penalty that is added to the weight. The examiner further notes that weights are stored as bits which are different between different weights such as smaller and larger weights. The examiner further notes that YANG and XU are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s loss function to incorporate the loss function is configured to penalize a first type of bits more than a second type of bits in weights as taught by XU [0110] to further boost the model accuracy [0110]).
Regarding claim 7, YANG teaches the method of claim 6, however, YANG is not relied upon to explicitly teach wherein bits of the first type have a value of one; and bits of the second type have a value of zero On the other hand, XU teaches wherein bits of the first type have a value of one; and bits of the second type have a value of zero ([0110] For this new linear model, an L2 regularizer, such as a Ridge regularizer, is applied, with a weaker penalty strength compared to the A used in the MCP regression. The examiner notes that XU teaches a loss function that uses L2 regularization which penalizes larger weights more than smaller weights due to the squared penalty that is added to the weight. The examiner further notes that weights are stored as bits which are different between different weights such as smaller and larger weights, and those bits that represent a larger weight or a smaller weight can have values or zeros or ones. The examiner further notes that YANG and XU are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s loss function to incorporate wherein bits of the first type have a value of one; and bits of the second type have a value of zero as taught by XU [0110] to further boost the model accuracy [0110]).
Claim 14 is the computing device claim corresponding to the method claim 5, therefore, it is rejected based upon the same rationale as the rejection of claim 5.
Claim 15 is the computing device claim corresponding to the method claim 6, therefore, it is rejected based upon the same rationale as the rejection of claim 6.
Claim 19 is the non-transitory computer storage medium claim corresponding to the method claim 5, therefore, it is rejected based upon the same rationale as the rejection of claim 5.
Claim 20 is the non-transitory computer storage medium claim corresponding to the method claim 6, therefore, it is rejected based upon the same rationale as the rejection of claim 6.
Claims 10, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over YANG (Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruning), in view of YANG2 (Multi-Objective Pruning for CNNs Using Genetic Algorithm).
Regarding claim 10, YANG teaches wherein the re-training further includes determining an accuracy performance level of the re-trained weight matrix ([Page 7074, Sec. 4] Once each individual layer has been pruned using Step 2 to 4, Step 5 performs global fine-tuning of weights across the entire network using back-propagation as described in Sec. 4.5. All these steps are iteratively performed until the final network can no longer maintain a given accuracy, e.g., 1% accuracy loss.)
determining an energy performance level of the re-trained weight matrix ([Page 6075, Sec. 4.1] At the beginning of each outer loop iteration in Fig. 2, the new pruning order is redetermined according to the new energy estimation of each layer.)
however, YANG is not relied upon to explicitly teach:
evaluating a combined performed level based on the accuracy performance level and the energy performance level.
Searching for a weight selection and modification solution to improve or optimize the combined performed level.
On the other hand, YANG2 teaches evaluating a combined performed level based on the accuracy performance level and the energy performance level ([Page 1-2, Sec. 2.1] Similar to general evolutionary algorithms, we design a fitness function f to evaluate the comprehensive performance of a genome. In our method, f is defined by the weighted average of error rate e, computation remained rate c and sparsity s. And our target is to minimize the fitness function as follows:
PNG
media_image2.png
78
662
media_image2.png
Greyscale
The coefficients λ1, λ2, and λ3 adjust the importance of the three objectives.
e, c and s denote the percentage of misclassified samples, remained multiplication-
addition operations (FLOPs) and zeroed out parameters, respectively. The examiner notes that YANG2 teaches the evaluation of a fitness function that takes into account the error rate (performance), and FLOPS (energy consumption). The examiner further notes that YANG and YANG2 are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s machine learning method to incorporate evaluating a combined performed level based on the accuracy performance level and the energy performance level as taught by YANG2 [page 1-2, Sec. 2.1] to minimize computational cost while maintaining tiny accuracy loss [abstract]).
Furthermore, YANG2 teaches Searching for a weight selection and modification solution to improve or optimize the combined performed level ([Page 4, Main Procedure] After each heuristic pruning process including selection, crossover and mutation with T iterations, we retrain the elite genome so that the remained weights can compensate for the loss of accuracy, and then reinitialize the population by the trained elite genome. The above procedures are repeated iteratively until the fitness of the elite genome is convergence. The examiner notes that YANG and YANG2 are both directed to machine learning and both are reasonably analogous to the claimed invention. Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified YANG’s machine learning method to incorporate Searching for a weight selection and modification solution to improve or optimize the combined performed level as taught by YANG2 [Page 4, Main Procedure] in order for the fitness of the elite genome to converge [Page 4, Main Procedure]).
Claim 17 is the non-transitory computer storage medium claim corresponding to the method claims 2 and 10, therefore, it is rejected based upon the same rationale as the rejection of claims 2 and 10.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
MORALIS (Neuromorphic Silicon Photonics and Hardware-Aware Deep Learning for High-Speed Inference)
“MORALIS teaches a review of recent progress in integrated photonic neuromorphic architectures and analyzes the architectural and photonic hardware-based factors that limit their performance”
SORBARO (Optimizing the Energy Consumption of Spiking Neural Networks for Neuromorphic Applications)
“SORBARO teaches that quantization-aware training of CNNs leads to better accuracy in SNNs”
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHAMCY ALGHAZZY whose telephone number is (571)272-8824. The examiner can normally be reached on M-F 7:30am-5:00pm EST.
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, OMAR FERNANDEZ RIVAS can be reached on (571) 272-2589. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/SHAMCY ALGHAZZY/Examiner, Art Unit 2128
/OMAR F FERNANDEZ RIVAS/Supervisory Patent Examiner, Art Unit 2128
Prosecution Insights