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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on April 17, 2026 has been entered.
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 1-3, 6-12 and 15-24 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
With respect to claim 1”
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites mental process.
A method for managing machine learning features, the method comprising:
Tracking access, wherein the access are made by a machine learning system to feature types of a set of features, wherein the tracking generates an access count for each of the individual feature types; Tracking access and generating a count for each feature could be practically performed in the mind.
Generating ranks for the individual feature types of the set of features based on the access count, wherein higher rank is associated with more accesses and lower rank is associated with fewer access; and Generating a ranking for the features based on the count could be practically performed in the mind.
Assigning the individual features to different levels of a memory hierarchy based on the rank, wherein higher ranked feature types are assigned to levels of the memory hierarchy with lower latency and lower ranked features are assigned to levels of the memory hierarchy with lower latency and lower ranked feature types are assigned to levels of the memory hierarchy with higher latency. Assigning features to different levels of memory based on a ranking could be practically performed in the mind.
Step 2A, Prong 2:
No additional elements are recited.
Step 2B:
No additional elements are recited.
The claim is ineligible.
With respect to claim 2:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Applying a weight to the access count to generate a weighted access count. Applying a weight to generate a weighted access count could be practically performed in the mind.
With respect to claim 3:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Generating the ranks occurs based on the weighted access count. Generating ranks could be practically performed in the mind.
With respect to claim 6:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Generating new features based on the set of features. Generating new features could be practically performed in the mind.
With respect to claim 7:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Filtering the new features. Filtering features could be practically performed in the mind.
With respect to claim 8:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Generating a score from the set of features. Generating a score could be practically performed in the mind.
With respect to claim 9:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Generating comprises performing one or both of crossing and discretization on the set of features. Discretization of a set of features is a mathematical operation.
Claims 10-12, 15-18 and 19-24 are rejected according to claims 1-3 and 6-9.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 10 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Roberts et al. (US PG Pub 2023/0051103) and Mavrommatis et al (US PG Pub 2021/0374127).
With respect to claim 1:
Roberts teaches:
A method for managing machine learning features, the method comprising:
Tracking access, wherein the tracking generates an access count for each of the individual feature types; (Paragraph [044], discloses monitoring access counts for each block of data and generating an access count for each block of data)
Generating ranks for the individual feature types of the set of features based on the access count, wherein higher rank is associated with more accesses and lower rank is associated with fewer access; and (Paragraph [044], discloses ranking the blocks of data based on the number of accesses, the ranks being cold data or hot data)
Assigning the individual features to different levels of a memory hierarchy based on the rank, wherein higher ranked feature types are assigned to levels of the memory hierarchy with lower latency and lower ranked features are assigned to levels of the memory hierarchy with lower latency and lower ranked feature types are assigned to levels of the memory hierarchy with higher latency. (Paragraph [044], discloses assigning data blocks to memory based on their ranking, wherein the hot data (higher ranked) is assigned to lower latency memory and the cold data is assigned to slower memory)
Roberts does not appear to explicitly disclose:
Wherein the access are made by a machine learning system to feature types of a set of features.
Mavrommatis teaches:
Wherein the access are made by a machine learning system to feature types of a set of features. (Paragraphs [022] and [023], discloses stateful features, which are types of features and include access counts, etc. Said stateful features are then accessed to train and implement learning models)
It would have been obvious for one of ordinary skill in the art before the effective filing date of the present application to implement a method that utilized the teachings of Roberts and the teachings of Mavrommatis, both in the same field of invention. This would allow for a more efficient method to manage the complete lifecycle of feature data in a scalable and reusable manner (Mavrommatis Paragraph [005]).
Claims 10 and 19 are rejected according to the rejection of claim 1.
Claims 2, 3, 11, 12, 20 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Roberts et al. (US PG Pub 2023/0051103) in view of Mavrommatis et al (US PG Pub 2021/0374127) and Richardson et al (US Patent 7,783,632).
With respect to claim 2:
The combination of Roberts and Mavrommatis does not appear to explicitly disclose:
Applying a weight to the access count to generate a weighted access count.
Richardson teaches:
Applying a weight to the access count to generate a weighted access count. (Column 1 Line 57, “In general, the popularity based ranking of an object can be determined in whole or in part by counting the number of times the object is accessed. The weight of the count can be affected by other factors such as the user's action performed with respect to the object, the rate at which the object is accessed by the same user or by different users, or the user ID or machine ID that accessed the object.” This discloses that the access count weights differently.)
Accordingly, it would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, having the combination of Roberts and Mavrommatis and Richardson before them, specifically , to incorporate Richardson’s teachings of ranking objects/features by weights access count. One would have been motivated to make such combination in order to improve the efficiency of ranking objects/features and the importance of each feature/object that is ranked
With respect to claim 3:
Richardson teaches:
Generating the rank occurs based on the weighted access count. (Richardson, Column 2 Line 31, “The various features can be weighted depending on whether a specific ranking is desired. For example, if the user would like to only rank objects accessed in the morning hours, then the training data used to teach the ranking component can rely more heavily on or at least include the time the object is accessed. Thus, the access time feature can be given greater weight than some of the other features. Conversely, if time is less important to this ranking scheme, then the time feature can be weighted less than the other features or given no weight at all.” This teaches that the rankings consider the weightage of each feature’s access count.)
Claims 11, 12 and 20 and 21 are rejected according to the rejections of claim 2 and 3.
Claims 6, 7, 9, 15, 16, 18, 22 and 23 are rejected under 35 U.S.C. 103 as being unpatentable over Roberts et al. (US PG Pub 2023/0051103) in view of Mavrommatis et al (US PG Pub 2021/0374127) and Li et al (CN 113627422).
With respect to claim 6:
The combination of Roberts and Mavrommatis does not appear to explicitly disclose:
Generating new features based on the set of features.
Li teaches:
Generating new features based on the set of features. (Page 10, “the model to be trained is used to perform feature fusion processing on the first feature and the second feature to obtain the third feature.” This teaches that a third feature was generated using a set of features that contains first feature and second feature.)
Accordingly, it would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, having the combination of Roberts, Mavrommatis and Li before them, specifically , to incorporate Li’s teachings of creating a new feature based on a set of existing features. One would have been motivated to make such combination in order to improve the set of features for better training of a machine learning model.
With respect to claim 7:
Roberts teaches:
Filtering the new features. (Paragraph [049], discloses filtering data blocks)
With respect to claim 9:
Li teaches:
Generating the new features comprises performing one or both of crossing and discretization on the set of features. (Page 10, “the feature fusion processing includes at least one of addition processing, multiplication processing, subtraction processing, cascade processing, and cascade convolution processing.” This discloses that a new feature being generating using multiple process that includes addition(combing) two features.)
Claims 15, 16, 18, 22 and 23 are rejected according to the rejections of claim 6, 7 and 9.
Claims 8 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Roberts et al. (US PG Pub 2023/0051103) in view of Mavrommatis et al (US PG Pub 2021/0374127) and Shaked et al (US Patent 9,805,312)
With respect to claim 8:
The combination of Roberts and Mavrommatis does not appear to explicitly disclose:
Generating a score from the set of features.
Shaked teaches:
Generating a score from the set of features. (Column 6 Line 37, “The score for each ranking criterion may be combined to generate a total score for each feature. Accordingly, all the features in a template may be ranked based on the total score generated for each feature.” Column 7 Line 37, “Based on the total scores for each feature shown in Table 1 above, it may be determined that a first set of features exceeds a threshold ranking criteria, e.g., any one or combination of a total score of 1500, a number of occurrences of 950, a number of impressions of 600, and the like.” This discloses that each feature has a score and the combination of all feature in a set of features will have a total score.)
It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, having the combination of Roberts and Mavrommatis before them, specifically, it would have been obvious to score the features of a machine learning model based on its usage or accessed by user. The rationale for this combination lies in improving the efficiency of accessing the important features that are stored in lower memory levels and improving the performances of training a machine learning model with such process.
Claim 18 is rejected according to the rejection of claim 8.
Response to Arguments
Claim Rejections - 35 USC § 101
Applicant argues “Specifically, the claimed techniques provide an improvement in feature management for a computer-based machine learning system. In particular, access to different individual features of a set of features are tracked. Based on this tracking, features are ranked, with a higher rank being associated with greater accesses and a lower rank being associated with fewer accesses. Features are assigned to different levels of a memory hierarchy based on rank. These operations represent an improvement to computer operations.”
Examiner respectfully disagrees. MPEP 2106.05(a)(II) recites “However, it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology.” The instant limitations have all been identified as mental process as stated above in the body of the rejection, as stated in the rejection there are no additional elements recited in the claim.
Claim Rejections - 35 USC § 103
Applicant’s arguments with respect to the 35 USC 103 rejections have been considered but are moot because the new ground of rejection.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MARIELA D REYES whose telephone number is (571)270-1006. The examiner can normally be reached Monday-Friday, 7:30 am -5:00 pm.
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, David Wiley can be reached at (571) 272-3923. 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.
/Mariela Reyes/Supervisory Patent Examiner, Art Unit 2142
Prosecution Insights