DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This office action is in response to submission of application on 3/24/2023.
Claims 1-20 are presented for examination.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Is the claim to a process, machine, manufacture or composition of matter?
Claim 1-7 are directed to a method; claims 8-14 are directed to a system; and claims 15-20 are directed to a non-transitory computer-readable medium; therefore, all claims are directed to one of the four statutory categories.
Step 2A Prong One: Does the claim recite an abstract idea, law of nature, or natural
phenomenon?
Claim 1 recites limitations of:
determining whether the first metric value satisfies a first predetermined condition; - mental process (observation, evaluation, judgement) as a human mind can evaluate whether a value meets a certain condition.
based on determining that the first metric value does not satisfy the first predetermined condition: determining not to train the obtained first subnetwork for the one-shot NAS; - mental process (observation, evaluation, judgement) as a human mind can evaluate whether to proceed if prior criteria is not met.
Step 2A Prong Two: Does the claim recite addition elements that integrate the judicial exception into a practical application?
Claim 1 recites additional elements of:
obtaining an overall network, the overall network comprising a plurality of candidate subnetworks for the one-shot NAS; - obtaining an overall network merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g), item (3), which identifies necessary data gathering and outputting as an example of extra-solution activity.
obtaining a first subnetwork of the plurality of candidate subnetworks from the overall network; - obtaining a first subnetwork from the overall network merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g), item (3), which identifies necessary data gathering and outputting as an example of extra-solution activity.
obtaining a first metric value of the first subnetwork; - obtaining a first metric value merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g), item (3), which identifies necessary data gathering and outputting as an example of extra-solution activity.
and obtaining a second subnetwork of the plurality of candidate subnetworks from the overall network; - obtaining a second subnetwork from the overall network merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g), item (3), which identifies necessary data gathering and outputting as an example of extra-solution activity.
and training the second subnetwork for the one-shot NAS. - this element constitutes “mere instructions to apply an exception.” (MPEP § 2106.05(f)). This limitation is using the abstract idea, of determining whether values meets a predetermined condition, and apply to “training the second subnetwork for the one-shot NAS”.
The additional elements do not integrate the abstract idea into a practical application.
Step 2B: Does the claim recite additional elements that amount to significantly more than the
judicial exception?
The additional elements are:
obtaining an overall network, the overall network comprising a plurality of candidate subnetworks for the one-shot NAS; - obtaining an overall network merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g). Data gathering is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(iv).
obtaining a first subnetwork of the plurality of candidate subnetworks from the overall network; - obtaining a first subnetwork from the overall network merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g). Data gathering is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(iv).
obtaining a first metric value of the first subnetwork; - obtaining a first metric value merely amounts to data gathering which is insignificant extra-solution activity See MPEP § 2106.05(g). Data gathering is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(iv).
and obtaining a second subnetwork of the plurality of candidate subnetworks from the overall network; - obtaining a second subnetwork from the overall network merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g). Data gathering is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(iv).
and training the second subnetwork for the one-shot NAS. - this element constitutes “mere instructions to apply an exception.” (MPEP § 2106.05(f)). This limitation is using the abstract idea, of determining whether values meets a predetermined condition, and apply to “training the second subnetwork for the one-shot NAS”.
The additional elements do not amount to significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Independent claim 8 and 15 recites the same relevant limitations and a similar analysis applies. Claim 8 recites the additional elements of, “A system for performing a one-shot neural architecture search (NAS), the system comprising: at least one memory storing instructions; and at least one processor configured to execute the instructions to:” – components recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2). Claim 15 recites additional elements of, “A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause the at least one processor to:” – components recited at a high level are construed as generic computer components used to implement the abstract idea. See MPEP 2106.05(f)(2). They do not integrate the abstract idea into a practical application. Not do they amount to significantly more. Therefore, the independent claims are not patent eligible.
The above analysis similarly applies to the dependent claims.
Dependent claim 2, 9, and 16 recites,
“obtaining a second metric value of the second subnetwork; - obtaining a second metric value is merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g). Data gathering is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(iv),
and determining whether the second metric value satisfies the first predetermined condition;
- mental process (observation, evaluation, judgement) as a human mind is able to determine if the second metric satisfies a condition,
and wherein the second subnetwork for the one-shot NAS is trained based on determining that the second metric value satisfies the first predetermined condition.” - this element constitutes “mere instructions to apply an exception.” (MPEP § 2106.05(f)). This limitation is using the abstract idea, “determining that the second metric value satisfies the first predetermined condition” in order to train the second subnetwork for the one-shot NAS.
Dependent claim 3, 10, and 17 recites, “comparing the first metric value to a first predetermined threshold.” – mental process (observation, evaluation, judgement) as a human mind is able to compare a value to a threshold.
Dependent claim 4, 11, and 18 recites, “determining whether the first metric value is greater than a first predetermined threshold and less than a second predetermined threshold.” – mental process as the human mind can determine if a metric value is greater than a threshold and less than another threshold.
Dependent claim 5, 12, and 19 recites, “determining whether the plurality of metric values respectively satisfy a corresponding plurality of predetermined conditions, including the first predetermined condition, for training in the one-shot NAS.” – mental process (observation, evaluation, judgement) as a human mind can observe whether the values satisfy predetermined conditions.
Dependent claim 6, 13, and 20 recites, “first metric value comprises at least one of a latency value, a model size value, and a floating point operations per second (FLOPS) value.” – a metric value that could be one of these three attributes merely amounts to data gathering which is insignificant extra-solution activity. See MPEP § 2106.05(g). Data gathering is well-understood, routine, and conventional. See MPEP 2106.05(d)(II)(iv),
Dependent claim 7 and 14 recites, “first predetermined condition is determined based on at least one of an overall number of layers in a subnetwork, an overall number of convolutional layers in a subnetwork, an overall number of residual blocks in a subnetwork, and an overall number of transport layers in a subnetwork.” – mental process (observation, evaluation, judgement) as a human mind can observe whether the first predetermined condition is based on one of the four options.
The dependent claims do not integrate the abstract idea into a practical application, not do they amount to significantly more than the abstract idea.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Peng et al. (Cream of the Crop: Distilling Prioritized Paths For One-Shot Neural Architecture Search, herein Peng) in view of Lok Won Kim (US 20230090720 A1, Optimization for Artificial Neural Network Model and Neural Processing Unit, herein Kim).
Regarding claim 1,
Peng teaches:
obtaining an overall network, the overall network comprising a plurality of candidate subnetworks for the one-shot NAS; (Peng, page 2, section 2, "The architecture search space A is encoded in a hypernetwork, denoted as N(A,W), where W is the weight of the hypernetwork. The weight W is shared across all the architecture candidates, i.e., subnetworks α ∈ A")
obtaining a first subnetwork of the plurality of candidate subnetworks from the overall network; (Peng, page 3, section 3, "for each batch, we randomly sample a single path α from the hypernetwork N")
obtaining a first metric value of the first subnetwork; (Peng, page 4, algorithm 1, line 7, "Calculate Flops(α) and top-1 accuracy Accα on val subset", note: for every path, FLOPS and accuracy is computed. FLOPS is considered one of the metric values, therefore since there is a flops that is computed, it means that is has been obtained.)
determining whether the first metric value satisfies a first predetermined condition; (Peng, page 3, section 3.1, "Accval((N(α,wα)) ≥ Accval(N(ˆαk,wˆαk)) & Flops(α) ≤ Flops(ˆαk), ˆαk ∈ B", note: this is an explicit conditional check on the metric values of the sampled path. The FLOPS is compared against threshold values)
Peng does not explicitly teach,
based on determining that the first metric value does not satisfy the first predetermined condition:
determining not to train the obtained first subnetwork for the one-shot NAS; and
obtaining a second subnetwork of the plurality of candidate subnetworks from the overall network; and
training the second subnetwork for the one-shot NAS.
Kim teaches,
based on determining that the first metric value does not satisfy the first predetermined condition: determining not to train the obtained first subnetwork for the one-shot NAS; (Kim, paragraph [0181], " Next, a training process may be performed on a new BNN model having the adjusted number of channels. If the accuracy matches the threshold, the model can be selected as the best BNN model with a given initial depth (i.e., number of layers). If the accuracy is less than the threshold, the number of channels in each layer may be increased by one, and the model may be retrained to check whether the accuracy is improved. This process can be repeated until an optimal model with accuracy matching the threshold value is found", note: First metric value maps to accuracy, subnetworks maps to the models, and predetermined condition maps to below threshold. When the accuracy of the model is below the threshold, the BNN weight is adjusted and retrained. However, when it is above, it is not retrained because it does not satisfy the threshold. Therefore, the subnetworks that do not satisfy the predetermined condition is not trained.)
obtaining a second subnetwork of the plurality of candidate subnetworks from the overall network; and training the second subnetwork for the one-shot NAS. (Kim, paragraph [0181], " This process can be repeated until an optimal model with accuracy matching the threshold value is found", note: when the model fails to meet an accuracy threshold, the model is retrained and this process is repeated until a model meeting the threshold is found. This disclosure teaches obtaining a new candidate model and training it after a prior model fails to satisfy the threshold. Therefore, there are subsequent models selected for training which corresponds to subsequent (second) subnetworks.)
It would be obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teaching of Peng and Kim because Kim demonstrates how to use the data from Peng to filter out subnetworks that do not meet the threshold/condition. By pre-filtering, unnecessary training is avoided, saving time and cost by ensuring training steps are only spent on viable candidates.
Regarding claim 2,
The combination of Peng and Kim teaches, The method of claim 1, wherein the training the second subnetwork comprises:
obtaining a second metric value of the second subnetwork; (Peng, algorithm 1, line 7, "Calculate Flops(α) and top-1 accuracy Accα on val subset", note: for every path, FLOPS and accuracy is computed before any training decision is made. FLOPS is one of the metric values, therefore, FLOPS maps to metric value)
and determining whether the second metric value satisfies the first predetermined condition; and wherein the second subnetwork for the one-shot NAS is trained based on determining that the second metric value satisfies the first predetermined condition. (Peng, page 4, section 3, " If both performance and complexity satisfy α Eq. (3), then the path α is added into the prioritized path board B." note: this states that if the path satisfies constraints and conditions, then it is proceeded into training.)
Regarding claim 3,
The combination of Peng and Kim teaches, The method of claim 1, wherein the determining whether the first metric value satisfies the first predetermined condition comprises comparing the first metric value to a first predetermined threshold. (Peng, page 4, algorithm 1, line 1, " Random initialize W, θ, B with path Flops ∈ [min,max]", note: this is a distinct threshold that the first metric value has to satisfy.)
Regarding claim 4,
The combination of Peng and Kim teaches, The method of claim 1, wherein the determining whether the first metric value satisfies the first predetermined condition comprises determining whether the first metric value is greater than a first predetermined threshold and less than a second predetermined threshold. (Peng, page 4, algorithm 1, line 1, " Random initialize W, θ, B with path Flops ∈ [min,max]", note: the [min, max] flops intervale is a two sided bound with an explicit lower threshold (min) and upper threshold (max). Therefore, this would map to first metric value is greater than first predetermined threshold and less than a second predetermined threshold)
Regarding claim 5,
The combination of Peng and Kim teaches, The method of claim 1, wherein the obtaining the first metric value comprises obtaining a plurality of metric values, including the first metric value of the first subnetwork; and wherein the determining whether the first metric value satisfies the first predetermined condition comprises determining whether the plurality of metric values respectively satisfy a corresponding plurality of predetermined conditions, including the first predetermined condition, for training in the one-shot NAS. (Peng, page 4, algorithm 1, line 7, "Calculate Flops(α) and top-1 accuracy Accα on val subset", note: there are two distinct metric values that are computed for every sampled path in the same step. Flops and accuracy are the plurality of metric values.)
Regarding claim 6,
The combination of Peng and Kim teaches, The method of claim 1, wherein the first metric value comprises at least one of a latency value, a model size value, and a floating point operations per second (FLOPS) value. (Peng, page 4, algorithm 1, line 7, "Calculate Flops(α) and top-1 accuracy Accα on val subset", note: flops maps to floating point operations per second)
Regarding claim 7,
The combination of Peng and Kim teaches, The method of claim 1, wherein the first predetermined condition is determined based on at least one of an overall number of layers in a subnetwork, an overall number of convolutional layers in a subnetwork, an overall number of residual blocks in a subnetwork, and an overall number of transport layers in a subnetwork. (Peng, page 7, section 4, "we evaluate our method on more challenging spaces, i.e., the combinations of operators from different designed space, including, MBConv [10], Residual Block [12] and normal 2D convolutions.")
Claims 8-14 is a system claim, A system for performing a one-shot neural architecture search (NAS), the system comprising: at least one memory storing instructions; and at least one processor configured to execute the instructions to: (Kim, paragraph [0019], "The computing system may include at least one processor; and at least one memory, operably electrically connected to the at least one processor, configured to store an instruction for a computer-implemented apparatus that searches an optimal design of an artificial neural network (ANN)"), that corresponds to method claims 1-7. Otherwise, they are not patentably distinguishable. Therefore, claims 8-14 are rejected for the same reasons as claims 1-7, respectively.
Claims 15-20 is a machine, A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause the at least one processor to: (Kim, paragraph [0020], "According to an example of the present disclosure, a non-transitory computer-readable storage medium storing instructions for a computer-implemented apparatus that searches an optimal design of an artificial neural network (ANN)"), that corresponds to method claims 1-6. Otherwise, they are not patentably distinguishable. Therefore, claims 15-20 are rejected for the same reasons as claims 1-7, respectively.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to UYEN-NHU PHAM TRAN whose telephone number is (571)272-1559. The examiner can normally be reached Monday - Friday 7:30-5.
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, Miranda Huang can be reached at (571) 270-7092. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/U.P.T./Examiner, Art Unit 2124
/MIRANDA M HUANG/Supervisory Patent Examiner, Art Unit 2124