Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
DETAILED ACTION
Status of the Application
This action is in response to the Amendment filed on 7/16/2026, and is a Final Office Action. Claims 1-4, 6-10 are pending in the application.
Claim Analysis - 35 USC § 101
Claims 1-4, 6-10 are eligible under 35 USC 101. Under Step 1 of the analysis, the pending independent claims 1, 7, 9 are directed to a device, method and recording medium, respectively, thus
meeting the Step 1 eligibility criterion. Under Prong One of Revised Step 2A of the 2019 PEG, the pending independent claims do not recite a judicial exception. Thus, pending claims 1-4, 6-10 are directed to patent-eligible subject matter.
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.
Claims 1, 2, 3, 4, 6, 7, 8, 9, 10 are rejected under 35 U.S.C. 103 as being unpatentable in view of Brothers (20160358070) in further view of Yang (KR20220090175A).
As per Claims 1, 7, 9, Brothers teaches a device, method and CRM comprising:
A memory configured to store instructions; at least one processor configured to execute the instructions to: (the memory/processor represent generic computing elements that perform the claimed limitations. At least: para 6)
Train a neural network model with a first neural network structure using a data set for training;
(at least para 27)
Generate analysis information indicating an importance of each of elements comprising the first neural network structure by analyzing a trained model, wherein the trained model is generated from the neural network model by the training using the data set for training; (at least para 23: “Example embodiments include a framework that automatically tunes parameters of an input neural network and outputs a modified (e.g., optimized) neural network. Example embodiments further include a method of determining tuned parameters for a neural network. In one arrangement, opportunities for modifying a neural network may be identified. The neural network may be modified based upon the identified opportunities. Examples of different modifications that may be applied to a neural network include, but are not limited to, pruning, decomposition, precision and/or numerical format modification, convolution kernel substitution, activation function substitution, kernel fusion, and/or scaling. The modified neural network may be validated to determine whether the modified neural network meets established performance requirements.”)
Identify an element in the first neural network structure whose importance is lower than a predetermined value using the generated analysis information; (at least para 23, 26-29, 36-40, fig 1-3 and associated/related text; at least para 70-71)
Generate a second neural network structure based on the first neural network structure by deleting the identified element from the first neural network structure (at least para 23, 26-29, 36-40, fig 1-3 and associated/related text; at least para 70-71)
: each element comprises weights in a kernel used in a convolution layer that constitutes the first neural network structure; (at least para 4, 27, 43)
Yang further teaches:
The processor is further configured to execute the instructions to remove the kernel that comprise more than a predetermined percentage of the identified weights. (at least: page4: “According to an embodiment of the present invention, the aforementioned kernel deletion cycle may be performed based on at least one data set. That is, one processor110 can apply the kernel deletion cycle to all data sets. Therefore, even in the same neural network, when the kernel deletion cycle is applied based on different data sets,the remaining kernels and nodes may be different”, “Next, the processor 110 may determine the amount of change in the first distribution of weights for the kernel through pattern analysis after learning is finished [S420]. Next, the processor 110 may delete the kernel when the distribution change amount is equal to or less than or equal to a preset value [S430]. Next, the processor 110 may retrain the neural network from which the kernel has been deleted based on the training data [S440]. Next, after the re-learning is finished, the processor 110 may determine the amount of change in the second distribution of weights for the remaining kernel through pattern analysis [S450]. Next, the processor 110 may delete the kernel when the amount of change in the second distribution is equal to or less than or equal to a preset value [S460].”, abstract)
It would have been obvious for someone skilled in the art at the time of the filing of the
invention to modify Brothers’s existing features, with Yang’s feature of the processor is further configured to execute the instructions to remove the kernel that comprise more than a predetermined percentage of the identified weights, to reduce the weights of a neural network – Yang, abstract.
Furthermore, the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
As per Claims 2, 8, 10, Brothers in view of Yang teach :
Each element comprises a layer that constitutes the first neural network structure. (Brothers, at least: para 70-71)
As per Claim 3, Brothers in view of Yang teach :
Identifying the element in the first neural network structure comprises identifying a plurality of elements, and the processor is further configured to execute the instructions to: determine an element to be deleted among the identified plurality of elements using a processing time of an arithmetic unit in the layer. (at least: para 74: “ pruning may include reducing the precision of a numerical format of one or more numerical values of the neural network. For example, the neural network analyzer can analyze one or more weights in a number of layers of the neural network to determine whether the precision of the numerical format used for the weights may be reduced. By reducing the precision of numerical formats used, lower precision arithmetic hardware may in turn be used. Lower precision arithmetic hardware can be more power efficient and can be built more densely than higher precision arithmetic hardware. For purposes of illustration, a neural network that uses the minimum number of bits required to represent the range and precision of the parameters can achieve higher performance (e.g., faster runtime and/or lower power consumption) than a neural network using a greater number of bits than required.”)
As per Claim 4, Brothers in view of Yang teach :
Identifying the element in the first neural network structure comprises identifying a plurality of elements; and determine an element to be deleted among the identified plurality of elements using an execution efficiency of an arithmetic unit in the layer. (at least: para 6, 22, 27; para 74)
As per Claim 6, Brothers discloses :
Compute the importance based on an activation rate obtained by inputting data in the data set for training into the trained model. (at least: para 23, 42, 65)
Response to Arguments
Applicant’s arguments have been fully considered; Applicant argues with substance:
Applicant’s arguments have been considered but are moot in view of the new grounds of rejection.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALEXANDRU CIRNU whose telephone number is (571)272-7775. The examiner can normally be reached on M-F 9:00am-5pm. If attempts to reach the examiner by telephone are unsuccessful, the examiner's supervisor, Ilana Spar can be reached on (571) 270-7537. The fax phone number for the organization where this application or proceeding is assigned is 571- 273-8300. Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
Sincerely,
/Alexandru Cirnu/
Primary Patent Examiner, Art Unit 3622
8/20/2026
Prosecution Insights