DETAILED ACTION
This action is in response to the Applicant Response filed 16 April 2026 for application 17/966,568 filed 14 October 2022.
Claim(s) 1, 10, 17 is/are currently amended.
Claim(s) 4, 13, 20 is/are cancelled.
Claim(s) 1-3, 5-12, 14-19, 21-23 is/are pending.
Claim(s) 1-3, 5-12, 14-19, 21-23 is/are rejected.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant’s arguments regarding the 35 U.S.C. 101 rejection of claims 1-3, 5-12, 14-19, 21-23 have been fully considered but are not persuasive. Applicant argues that training is not a mental process and therefore creates eligibility. While Examiner agrees that training is not a mental process, this fact alone does not create eligibility. As noted in detail below, generic training of a model selection amounts to mere instructions to apply the exception (MPEP 2106.05(f)). Therefore, the 35 U.S.C. 101 rejection of claims 1-3, 5-12, 14-19, 21-23 is maintained.
Applicant’s arguments regarding the 35 U.S.C. 102 and/or 35 U.S.C. 103 rejections of the claims are based on the newly amended subject matter. All arguments are addressed in the 35 U.S.C. 102 and/or 35 U.S.C. 103 rejections of the claims below.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claim(s) 1-3, 5-12, 14-19, 21-23 is/are rejected under 35 U.S.C. 101, because the claim(s) is/are directed to an abstract idea, and because the claim elements, whether considered individually or in combination, do not amount to significantly more than the abstract idea, see Alice Corporation Pty. Ltd. V. CLS Bank International et al., 573 US 208 (2014).
Regarding claim 1, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 1 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of obtaining a set of training samples from one or more domains, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of weighting the output of each queried AI model using a respective weight of the plurality of weights that represents knowledge transferability for the corresponding queried AI model, thereby obtaining a plurality of weighted outputs, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of combining the weighted outputs to obtain a soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of adapting a target Al model via knowledge distillation using the soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – plurality of artificial-intelligence (Al) models, target Al model, knowledge distillation. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites training a model selector using the set of training samples to obtain a plurality of weights which is simply generic training to perform the abstract idea of model generation and amounts to mere instructions to apply the exception (MPEP 2106.05(f)).
The claim recites using the set of training samples to query a plurality of artificial-intelligence (Al) models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
generic training to perform the abstract idea and applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
plurality of artificial-intelligence (Al) models, target Al model, knowledge distillation amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 2, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 2 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method. The Step 2A Prong One Analysis for claim 1 is applicable here since claim 2 carries out the method of claim 1 but for the recitation of additional element(s) of using a transformer encoder for combining the weighted outputs.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – transformer encoder. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites using a transformer encoder for combining the weighted outputs which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
transformer encoder amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 3, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 3 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of obtaining the set of training samples from a plurality of domains, the set of training samples comprises a plurality of subsets of training samples obtained from the plurality of domains, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites using each subset of training samples to query the plurality of Al models except an excluded Al model of the plurality of Al models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
The claim recites wherein the excluded Al models excluded from said using each subset of training samples to query the plurality of AI models are different Al models which is simply additional information regarding the models, and the element(s) do(es) not apply the exception in a meaningful way (MPEP 2106.05(e)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
additional information regarding the models do(es) not apply the exception in a meaningful way (MPEP 2106.05(e))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 5, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 5 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of adapting the target Al model via the knowledge distillation based on Kullback-Leibler (KL) divergence of the output of the queried target Al model and the soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses performing KL divergence.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – Kullback-Leibler (KL) divergence. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites querying the target Al model using the set of training samples which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
Kullback-Leibler (KL) divergence amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 6, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 6 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of minimizing the KL divergence using a gradient decent method, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses performing gradient descent.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – gradient decent method. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
gradient decent method amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 7, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 7 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of evaluating a loss of the target Al model, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of updating a plurality of parameters based on the evaluated loss, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites wherein the plurality of parameters comprises one or more first parameters of the target Al model and a parameter used in said combining the weights outputs which is simply additional information regarding the parameters, and the element(s) do(es) not apply the exception in a meaningful way (MPEP 2106.05(e)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
additional information regarding the parameters do(es) not apply the exception in a meaningful way (MPEP 2106.05(e))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 8, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 8 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of evaluating a cross-entropy (CE) loss between the outputs of the queried target Al model and a set of labels corresponding to the set of query samples, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses calculating a loss.
The limitation of updating the plurality of parameters by minimizing the CE loss, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses minimizing a loss.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – cross-entropy (CE) loss. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites querying the target Al model using a set of query samples which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
cross-entropy (CE) loss amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 9, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 9 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) method.
The limitation of updating the plurality of parameters by minimizing the CE loss using a gradient decent method, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses performing gradient descent.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – gradient decent method. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
gradient decent method amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 10, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 10 is directed to an apparatus, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) apparatus.
The limitation of obtaining a set of training samples from one or more domains, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of weighting the output of each queried AI model using a respective weight of the plurality of weights that represents knowledge transferability from the corresponding queried AI model, thereby obtaining a plurality of weighted outputs, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of combining the weighted outputs to obtain a soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of adapting a target Al model via knowledge distillation using the soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – apparatus, at least one processor. The additional element(s) is/are recited at a high-level of generality (i.e., as generic computer components performing generic computer functions of executing instructions on the computers) such that it amounts to no more than mere instructions to apply the exception using generic computer components (MPEP 2106.05(b)).
The claim recites additional element(s) – plurality of Al models, target Al model, knowledge distillation. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites training a model selector using the set of training samples to obtain a plurality of weights which is simply generic training to perform the abstract idea of model generation and amounts to mere instructions to apply the exception (MPEP 2106.05(f)).
The claim recites using the set of training samples to query a plurality of artificial-intelligence (Al) models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
apparatus, at least one processor amount(s) to no more than mere instructions to apply the exception using generic computer components (MPEP 2106.05(b))
generic training to perform the abstract idea and applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
plurality of Al models, target Al model, knowledge distillation amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 11, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 11 is directed to an apparatus, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) apparatus. The Step 2A Prong One Analysis for claim 10 is applicable here since claim 11 carries out the apparatus of claim 10 but for the recitation of additional element(s) of using a transformer encoder for combining the weighted outputs.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – transformer encoder. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites using a transformer encoder for combining the outputs of the queried Al models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
transformer encoder amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 12, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 12 is directed to an apparatus, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) apparatus.
The limitation of obtaining the set of training samples from a plurality of domains, the set of training samples comprises a plurality of subsets of training samples obtained from the plurality of domains, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites using each subset of training samples to query the plurality of Al models except an excluded Al model of the plurality of Al models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
The claim recites wherein the excluded Al models excluded from said using each subset of training samples to query the plurality of AI models are different Al models which is simply additional information regarding the models, and the element(s) do(es) not apply the exception in a meaningful way (MPEP 2106.05(e)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
additional information regarding the models do(es) not apply the exception in a meaningful way (MPEP 2106.05(e))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 14, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 14 is directed to an apparatus, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) apparatus.
The limitation of adapting the target Al model via the knowledge distillation based on Kullback-Leibler (KL) divergence of the output of the queried target Al model and the soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses performing KL divergence.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) –KL divergence. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites querying the target Al model using the set of training samples which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
KL divergence amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 15, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 15 is directed to an apparatus, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) apparatus.
The limitation of evaluating a loss of the target Al model, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of updating a plurality of parameters based on the evaluated loss, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites wherein the plurality of parameters comprises one or more first parameters of the target Al model and a parameter used in said combining the weighted outputs which is simply additional information regarding the parameters, and the element(s) do(es) not apply the exception in a meaningful way (MPEP 2106.05(e)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
additional information regarding the parameters do(es) not apply the exception in a meaningful way (MPEP 2106.05(e))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 16, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 16 is directed to an apparatus, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites a(n) apparatus.
The limitation of evaluating a cross-entropy (CE) loss between the outputs of the queried target Al model and a set of labels corresponding to the set of query samples, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses calculating a loss.
The limitation of updating the plurality of parameters by minimizing the CE loss, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses minimizing a loss.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – CE loss. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites querying the target Al model using a set of query samples which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
cross-entropy (CE) loss amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 17, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 17 is directed to one or more computer-readable storage devices, which is directed to an article of manufacture, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites one or more computer-readable storage devices.
The limitation of obtaining a set of training samples from one or more domains, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of weighting the output of each queried AI model using a respective weight of the plurality of weights that represents knowledge transferability from the corresponding queried AI model, thereby obtaining a plurality of weighted outputs, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of combining the weighted outputs to obtain a soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of adapting a target Al model via knowledge distillation using the soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – one or more computer-readable storage devices, computer-executable instructions, processing structure. The additional element(s) is/are recited at a high-level of generality (i.e., as generic computer components performing generic computer functions of executing instructions on the computers) such that it amounts to no more than mere instructions to apply the exception using generic computer components (MPEP 2106.05(b)).
The claim recites additional element(s) – plurality of Al models, target Al model, knowledge distillation. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites training a model selector using the set of training samples to obtain a plurality of weights which is simply generic training to perform the abstract idea of model generation and amounts to mere instructions to apply the exception (MPEP 2106.05(f)).
The claim recites using the set of training samples to query a plurality of artificial-intelligence (Al) models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
one or more computer-readable storage devices, computer-executable instructions, processing structure amount(s) to no more than mere instructions to apply the exception using generic computer components (MPEP 2106.05(b))
generic training to perform the abstract idea and applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
plurality of Al models, target Al model, knowledge distillation amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 18, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 18 is directed to one or more computer-readable storage devices, which is directed to an article of manufacture, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites one or more computer-readable storage devices. The Step 2A Prong One Analysis for claim 17 is applicable here since claim 18 carries out the computer-readable storage device of claim 17 but for the recitation of additional element(s) of using a transformer encoder for combining the weighted outputs.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – transformer encoder. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites using a transformer encoder for combining the weighted outputs which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
transformer encoder amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 19, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 19 is directed to one or more computer-readable storage devices, which is directed to an article of manufacture, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites one or more computer-readable storage devices.
The limitation of obtaining the set of training samples from a plurality of domains, the set of training samples comprises a plurality of subsets of training samples obtained from the plurality of domains, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites using each subset of training samples to query the plurality of Al models except an excluded Al model of the plurality of Al models which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
The claim recites wherein the excluded Al models excluded from said using each subset of training samples to query the plurality of AI models are different Al models which is simply additional information regarding the models, and the element(s) do(es) not apply the exception in a meaningful way (MPEP 2106.05(e)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
additional information regarding the models do(es) not apply the exception in a meaningful way (MPEP 2106.05(e))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 21, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 21 is directed to one or more computer-readable storage devices, which is directed to an article of manufacture, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites one or more computer-readable storage devices.
The limitation of adapting the target Al model via the knowledge distillation based on Kullback-Leibler (KL) divergence of the output of the queried target Al model and the soft pseudo-label, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses performing KL divergence.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) –KL divergence. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites querying the target Al model using the set of training samples which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
KL divergence amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 22, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 22 is directed to one or more computer-readable storage devices, which is directed to an article of manufacture, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites one or more computer-readable storage devices.
The limitation of evaluating a loss of the target Al model, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
The limitation of updating a plurality of parameters based on the evaluated loss, as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process. The limitation is directed to observation, evaluation, judgment and opinion and is a process capable of being performed by a human mentally or using pen and paper.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind, then it falls within the "Mental Processes" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites wherein the plurality of parameters comprises one or more first parameters of the target Al model and a parameter used in said combining the weighted outputs which is simply additional information regarding the parameters, and the element(s) do(es) not apply the exception in a meaningful way (MPEP 2106.05(e)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
additional information regarding the parameters do(es) not apply the exception in a meaningful way (MPEP 2106.05(e))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Regarding claim 23, the claim is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 23 is directed to one or more computer-readable storage devices, which is directed to an article of manufacture, one of the statutory categories.
Step 2A Prong One Analysis: The claim recites one or more computer-readable storage devices.
The limitation of evaluating a cross-entropy (CE) loss between the outputs of the queried target Al model and a set of labels corresponding to the set of query samples, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses calculating a loss.
The limitation of updating the plurality of parameters by minimizing the CE loss, as drafted, is a process that, under its broadest reasonable interpretation, covers a mathematical concept. The limitation encompasses minimizing a loss.
If a claim limitation, under its broadest reasonable interpretation, covers performance of mathematical concepts, then it falls within the "Mathematical Concepts" grouping. Accordingly, the claim recites an abstract idea.
Step 2A Prong Two Analysis: With respect to the abstract idea, the judicial exception is not integrated into a practical application.
The claim recites additional element(s) – CE loss. The additional element(s) is/are recited at a high-level of generality such that it amounts to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h)).
The claim recites querying the target Al model using a set of query samples which is simply applying the models recited at a high level of generality and amounts to the recitation of the words “apply it” (or an equivalent) or amounts to no more than mere instructions to implement an abstract idea or other exception on a computer (MPEP 2106.05(f)).
Accordingly, the additional element(s) do(es) not integrate the abstract idea into a practical application because the additional element(s) do(es) not impose any meaningful limits on practicing the abstract idea, and, therefore, the claim is directed to an abstract idea.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element(s) of:
applying the models amount(s) to no more than mere instructions to apply the exception (MPEP 2106.05(f))
cross-entropy (CE) loss amount(s) to no more than indicating a field of use or technological environment in which to apply the judicial exception (MPEP 2106.05(h))
The additional element(s) do(es) not provide an inventive concept, and, therefore, the claim is not patent eligible.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1, 5, 7-8, 10, 14-17, 21-23 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mghabbar et al (Building a Multi-Domain Neural Machine Translation Model Using Knowledge Distillation, hereinafter referred to as “Mghabbar”) in view of Vinayagamurthy et al (US 2023/0058219 A1 - Secure Ensemble Training and Inference Using Heterogeneous Private Machine Learning Models, hereinafter referred to as "Vinayagamurthy").
Regarding claim 1 (Currently Amended), Mghabbar teaches a method comprising:
obtaining a set of training samples from one or more domains (Mghabbar, section 4.1 – teaches sentences from different domains; see also Mghabbar, Figures 1-2, section 4.2, section 5);
using the set of training samples to query a plurality of artificial-intelligence (Al) models (Mghabbar, sections 2, 4, 5 – teaches inputting multi-domain text data into difference teacher models; see also Mghabbar, Figure 2);
weighting the output of each queried AI model using a respective weight of the plurality of weights that represents knowledge transferability from the corresponding queried AI model, thereby obtaining a plurality of weighted outputs (Mghabbar, section 2.5 – teaches generating pseudo-labels from the outputs of the teacher models; Mghabbar, section 4.4 – teaches selecting the top-k teacher outputs to use for pseudo-label generation [selecting top-k is weighting the top-k]);
combining the weighted outputs to obtain a soft pseudo-label (Mghabbar, section 2.5 – teaches generating pseudo-labels from the outputs of the teacher models; Mghabbar, section 4.4 – teaches selecting the top-k teacher outputs to use for pseudo-label generation [selecting top-k is weighting the top-k]; see also Mghabbar, section 5, Figures 1-2); and
adapting a target Al model via knowledge distillation using the soft pseudo-label (Mghabbar, sections 2.5, 4 – teaches updating the student model using the pseudo-labels; see also Mghabbar, Figures 1-2).
However, Mghabbar does not explicitly teach a trainable model selector to weight the outputs of the queried AI models.
Vinayagamurthy teaches
training a model selector using the set of training samples to obtain a plurality of weights (Vinayagamurthy, [0020] – teaches using a meta-model to weight the outputs from the ensemble of models; Vinayagamurthy, [0026] – teaches training the meta-model based on meta-features; see also Vinayagamurthy, [0019] – extraction of meta-features using training data);
weighting the output of each queried AI model using a respective weight of the plurality of weights that represents knowledge transferability from the corresponding queried AI model, thereby obtaining a plurality of weighted outputs (Vinayagamurthy, [0020] – teaches weighting the outputs of the ensemble models using a meta-model).
It would have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to modify Mghabbar with the teachings of Vinayagamurthy in order to derive the importance that should be placed on each of the model responses in the field of ensemble based knowledge distillation (Vinayagamurthy, [0020] – “Using the meta-features from the data owners, the service provider generates a response to the query. In generating the response, the meta-features are evaluated using the meta-model to determine weights to be assigned to outputs on each of the individual data owner models. Thus, the purpose of the meta-model is to derive the importance that the query should place on the query response received from each of the data owners. The meta-model outputs weight vectors that are used by the query to place weights on the predictions or output provided by each local machine learning model within the ensemble. The weightings and which models are selected to be used in generating the response to the query are hidden by adding noise to the weight vectors. The final query response is generated by running an inner-product on the weight vector and the vector of the individual model outputs to hide the output of individual models.”).
Regarding claim 5 (Previously Presented), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 1 as noted above. Mghabbar further teaches wherein said adapting the target Al model via the knowledge distillation using the soft pseudo-label comprises:
querying the target Al model using the set of training samples (Mghabbar, sections 2.5, 4 – teaches inputting training samples into the student model); and
adapting the target Al model via the knowledge distillation based on Kullback-Leibler (KL) divergence of the output of the queried target Al model and the soft pseudo-label (Mghabbar, section 2.5 – teaches using KL divergence for updating the student model based on the model output and the pseudo-labels).
It would have been obvious to one of ordinary skill in the art before the filing data of the claimed invention to combine the teachings of Mghabbar and Vinayagamurthy for the same reasons as disclosed in claim 1 above.
Regarding claim 7 (Previously Presented), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 1 as noted above. Mghabbar further teaches
evaluating a loss of the target Al model (Mghabbar, section 2.5 – teaches determining a loss of the student model output and the pseudo-labels to update parameters); and
updating a plurality of parameters based on the evaluated loss (Mghabbar, section 2.5 – teaches determining a loss of the student model output and the pseudo-labels to update parameters);
wherein the plurality of parameters comprises one or more first parameters of the target Al model and a parameter used in said combining the weighted outputs (Mghabbar, sections 2.5, 4 – teaches using the loss function and knowledge distillation to update the student model as well as finetune the teacher models; Mghabbar, section 4.4 – teaches selecting the top-k teacher outputs to use for pseudo-label generation [selecting top-k is weighting the top-k]).
It would have been obvious to one of ordinary skill in the art before the filing data of the claimed invention to combine the teachings of Mghabbar and Vinayagamurthy for the same reasons as disclosed in claim 1 above.
Regarding claim 8 (Original), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 7 as noted above. Mghabbar further teaches
wherein said evaluating a loss of the target Al model comprises:
querying the target Al model using a set of query samples (Mghabbar, sections 2.5, 4 – teaches inputting training samples into the student model), and
evaluating a cross-entropy (CE) loss between the outputs of the queried target Al model and a set of labels corresponding to the set of query samples (Mghabbar, section 2.5 – teaches using cross-entropy to update the student model); and
wherein said updating the plurality of parameters based on the evaluated loss comprises: updating the plurality of parameters by minimizing the CE loss (Mghabbar, section 2.5 – teaches minimizing cross-entropy loss to update the student model).
It would have been obvious to one of ordinary skill in the art before the filing data of the claimed invention to combine the teachings of Mghabbar and Vinayagamurthy for the same reasons as disclosed in claim 7 above.
Regarding claim 10 (Currently Amended), it is the apparatus embodiment of claim 1 with similar limitations to claim 1 and is rejected using the same reasoning found in claim 1. Mghabbar further teaches an apparatus comprising:
at least one processor for performing actions (Mghabbar, section 5 – teaches running python code on GPUs) …
It would have been obvious to one of ordinary skill in the art before the filing data of the claimed invention to combine the teachings of Mghabbar and Vinayagamurthy for the same reasons as disclosed in claim 1 above.
Regarding claim 14 (Previously Presented), the rejection of claim 13 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy for the reasons set forth in the rejection of claim 5.
Regarding claim 15 (Previously Presented), the rejection of claim 10 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy for the reasons set forth in the rejection of claim 7.
Regarding claim 16 (Previously Presented), the rejection of claim 15 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy for the reasons set forth in the rejection of claim 8.
Regarding claim 17 (Currently Amended), it is the computer-readable storage device embodiment of claim 1 with similar limitations to claim 1 and is rejected using the same reasoning found in claim 1. Mghabbar further teaches one or more non-transitory computer-readable storage devices comprising computer-executable instructions, wherein the instructions, when executed, cause a processing structure to perform actions (Mghabbar, section 5 – teaches running python code on GPUs) …
It would have been obvious to one of ordinary skill in the art before the filing data of the claimed invention to combine the teachings of Mghabbar and Vinayagamurthy for the same reasons as disclosed in claim 1 above.
Regarding claim 21 (Previously Presented), the rejection of claim 20 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy for the reasons set forth in the rejection of claim 5.
Regarding claim 22 (Previously Presented), the rejection of claim 17 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy for the reasons set forth in the rejection of claim 7.
Regarding claim 23 (Previously Presented), the rejection of claim 22 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy for the reasons set forth in the rejection of claim 8.
Claim(s) 2, 11, 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mghabbar in view of Vinayagamurthy and further in view of Akbari et al. (VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text, hereinafter referred to as “Akbari”).
Regarding claim 2 (Previously Presented), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 1 as noted above. However, Mghabbar in view of Vinayagamurthy does not explicitly teach wherein said combining the outputs of the queried Al models comprises: using a transformer encoder for combining the outputs of the queried Al models.
Akbari teaches wherein said combining the outputs of the queried Al models comprises:
using a transformer encoder for combining the weighted outputs (Akbari, Figure 1, Sections 1, 3 – teaches a multiple models of differing modalities outputting data into a transformer encoder to combine the data for various downstream tasks).
It would have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to modify Mghabbar in view of Vinayagamurthy with the teachings of Akbari in order to provide transformers for effective learning of semantic representations and reduce the dependency on large-scale labeled data in the field of multi-domain knowledge distillation (Akbari, Section 5 – “In this paper, we present a self-supervised multimodal representation learning framework based on Transformers. Our study suggests that Transformers are effective for learning semantic video/audio/text representations — even if one model is shared across modalities — and multimodal self-supervised pre-training is promising for reducing their dependency on large-scale labeled data. We show that DropToken can significantly reduce the pre-training complexity with video and audio modalities and have minor impact on the models’ generalization. We report new records of results on video action recognition and audio event classification and competitive performance on image classification and video retrieval. Having these results, we still see some limitations in our work. Firstly, not all videos have organic audio or speech, while our approach depends on meaningful multimodal correspondences. Besides, the text modality currently consists of speech transcripts, which are noisy and sometimes sparse. Potential negative Societal Impacts are mainly concerned with applications. The models could be biased if one applies our approach to the multimodal videos that are not representative enough. Finally, our method is still demanding in computation, though we managed to avoid the need for human labels. Future work can improve upon these limitations.”).
Regarding claim 11 (Previously Presented), the rejection of claim 10 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy and further in view of Akbari for the reasons set forth in the rejection of claim 2.
Regarding claim 18 (Previously Presented), the rejection of claim 17 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy and further in view of Akbari for the reasons set forth in the rejection of claim 2.
Claim(s) 3, 12, 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mghabbar in view of Vinayagamurthy and further in view of Duerig et al. (US 2020/0401929 A1 – Systems and Methods for Performing Knowledge Distillation, hereinafter referred to as “Duerig”).
Regarding claim 3 (Previously Presented), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 1 as noted above. Mghabbar further teaches wherein said obtaining the set of training samples from the one or more domains comprises:
obtaining the set of training samples from a plurality of domains, the set of training samples comprises a plurality of subsets of training samples obtained from the plurality of domains (Mghabbar, section 4.2 – teaches generating subsets of domain specific training samples from the set of training samples).
However, Mghabbar in view of Vinayagamurthy does not explicitly teach wherein said using the set of training samples to query the plurality of Al models comprises: using each subset of training samples to query the plurality of Al models except an excluded Al model of the plurality of Al models; and wherein the excluded Al models of the plurality of subset of training samples are different Al models.
Duerig teaches wherein said using the set of training samples to query the plurality of Al models comprises:
using each subset of training samples to query the plurality of Al models except an excluded Al model of the plurality of Al models (Duerig, [0082]-[0083] – teaches selecting certain models based on the data [The selection of certain models is interpreted as similarly excluding certain models]); and
wherein the excluded Al models excluded from said using each subset of training samples to query the plurality of AI models are different Al models (Duerig, [0082]-[0083] – teaches selecting certain models based on a task [This is interpreted as excluding models that cannot perform the task and are, therefore, different models.]).
It would have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to modify Mghabbar in view of Vinayagamurthy with the teachings of Duerig in order to improve the performance of the distillation model by increasing the size and/or quality of the distillation training dataset in the field of multi-domain knowledge distillation (Duerig, [0027] – “In some implementations, the training examples in the initial dataset can include mostly weakly labeled or unlabeled training examples. By leveraging the knowledge contained in each machine-learned model, the outputs selected for inclusion in the distillation training dataset may be considered trusted without the need for a validation dataset (e.g., on a per-label or per-task basis). However, in certain implementations the initial training dataset can include or be associated with a validation dataset that includes labeled data (e.g., training examples that have a ground truth answer or label). For implementations that include a validation dataset, the validation dataset may be used to assist in the determination of an expert machine-learned model for performing a certain task. The designation of an expert machine-learned model can be used to assign high quality labels to the weakly labeled or unlabeled data that may improve the performance of the distillation model by increasing the size and/or quality of the distillation training dataset.”).
Regarding claim 12 (Previously Presented), the rejection of claim 10 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy and further in view of Duerig for the reasons set forth in the rejection of claim 3.
Regarding claim 19 (Previously Presented), the rejection of claim 17 is incorporated herein. Further, the limitations in this claim are taught by Mghabbar in view of Vinayagamurthy and further in view of Duerig for the reasons set forth in the rejection of claim 3.
Claim(s) 6, 9 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mghabbar in view of Vinayagamurthy and further in view of Kwon et al. (US 2021/0110273 A1 – Apparatus and Method with Model Training, hereinafter referred to as “Kwon”).
Regarding claim 6 (Original), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 5 as noted above. However, Mghabbar in view of Vinayagamurthy does not explicitly teach wherein said adapting the target Al model via the knowledge distillation based on the KL divergence of the output of the queried target Al model and the soft pseudo-label comprises: minimizing the KL divergence using a gradient decent method.
Kwon teaches wherein said adapting the target Al model via the knowledge distillation based on the KL divergence of the output of the queried target Al model and the soft pseudo-label comprises:
minimizing the KL divergence using a gradient decent method (Kwon, [0076] – teaches updating the target model parameters based on the gradient of the total loss function [including KL divergence as incorporated from Mghabbar]).
It would have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to modify Mghabbar in view of Vinayagamurthy with the teachings of Kwon in order to increase efficiency in the field of multi-domain knowledge distillation (Kwon, [0063] – “A lightening approach herein may simplify operations and provide a reduced size of the neural network, e.g., while maintaining accuracy of existing training models. Such a provided neural network may provide, for example, a reduction in latency, the protection of personal information, or a reduction in network traffic. The lightening may be classified into a lightweight approach that may maximize efficiency compared to existing models by providing an approach that is implemented with less computation, and may have a more efficient structure, and an approach that may use a scheme, such as a model compression to reduce parameters of an existing model.”).
Regarding claim 9 (Original), Mghabbar in view of Vinayagamurthy teaches all of the limitations of the method of claim 8 as noted above. However, Mghabbar in view of Vinayagamurthy does not explicitly teach wherein said updating the plurality of parameters by minimizing the CE loss comprises: updating the plurality of parameters by minimizing the CE loss using a gradient decent method.
Kwon teaches wherein said updating the plurality of parameters by minimizing the CE loss comprises:
updating the plurality of parameters by minimizing the CE loss using a gradient decent method (Kwon, [0076] – teaches updating the target model parameters based on the gradient of the total loss function [including cross-entropy as incorporated from Mghabbar]).
It would have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to modify Mghabbar in view of Vinayagamurthy with the teachings of Kwon in order to increase efficiency in the field of multi-domain knowledge distillation (Kwon, [0063] – “A lightening approach herein may simplify operations and provide a reduced size of the neural network, e.g., while maintaining accuracy of existing training models. Such a provided neural network may provide, for example, a reduction in latency, the protection of personal information, or a reduction in network traffic. The lightening may be classified into a lightweight approach that may maximize efficiency compared to existing models by providing an approach that is implemented with less computation, and may have a more efficient structure, and an approach that may use a scheme, such as a model compression to reduce parameters of an existing model.”).
Conclusion
Any inquiry concerning this communication or earlier communication from the examiner should be directed to MARSHALL WERNER whose telephone number is (469) 295-9143. The examiner can normally be reached on Monday – Thursday 7:30 AM – 4:30 PM ET.
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 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.
/MARSHALL L WERNER/ Primary Examiner, Art Unit 2125
Prosecution Insights