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 .
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 02 March 2026 has been entered.
Claims amended: 23
Claims cancelled: 1-22
Claims added: 24-27
Claims currently pending: 23-27
Response to Arguments
Applicant, in the “REMARKS” section and the “Summary of Amendments” sub-section, presents opening remarks regarding the disposition of the claims and the amendments to the claims. As no specific argument is raised in this/these section(s) with respect to the instant application, no rebuttal is required.
Applicant, in the “35 U.S.C. § 101 Rejections” section, refers to the previously presented grounds of rejection under said statute and requests withdrawal of the grounds of rejection. Unfortunately, grounds of rejection under this statute are presented herein. These grounds of rejection represent analysis under this statute that reflects the latest best practices and court analyses and they align with the analysis presented in the Appeal Decision dated 02 January 2026.
Applicant, in the “35 U.S.C. § 103 Rejections” section, refers to newly amended claim language in claim 23 and argues that Acuna Agost, the primary reference, does not teach “features characterizing an in-app event generated during execution of a mobile game”. Applicant then argues that Jiang “not disclose that the model updates are based on features characterizing in-app events generated during execution of the mobile game”. This line of argument is largely moot in view of the new grounds of rejection presented herein, which were necessitated by Applicant’s amendments to the claims. Examiner notes that Jiang does indeed teach “features characterizing in-app events generated during execution of the mobile game”. Jiang teaches computing a user response prediction for displaying ads, including tracking events related to those ads, and that the ads can be within a "gaming application". The advertisements are displayed “in-app” and the interactions with the advertisements are included in the features used in the determination of a likelihood of interaction. Therefore, Jiang, in combination with Acuna Agost, teaches this feature. Applicant then further asserts that Jiang “does not teach or suggest a two-stage training architecture” for machine learning training. This argument is misplaced as this feature of the claims is taught by Acuna Agost and Jiang is not relied upon to teach this feature. Examiner notes that claims 24-27 were newly added and any remarks made thereto are premature in light of the new grounds of rejection presented herein.
Applicant, on page 7, refers to Narsky and the newly amended claim language. Applicant asserts that Narsky does not cure the alleged deficiencies argued above. This argument is unpersuasive as Narsky is not relied upon herein to teach the features in question.
Applicant, on pages 7 and 8, argues that combining Acuna Agost with Narsky “would require a change in the fundamental operation mode” of the two references. Applicant argues that modifying Acuna Agost with the limited-memory BFGS of Narsky would require “substantial reconstruction of Acuna Agost’s learning architecture”. Examiner disagrees to this notion, as Acuna Agost, in at least 0063, discloses that the “machine learning module…to implement a generalized linear model”. Further, training of machine learning models is a known problem in the art. Swapping one known technique for another would have been obvious to try and would come with performance trade-offs that were known at the time the invention was filed. Applicant argues that swapping the “(FTRL)-proximal” algorithm of Acuna Agost with the limited-memory BFGS of Narsky, “would fundamentally alter the manner in which the model is trained and updated”. Fundamentally, the machine learning model training algorithm optimizes the weights of the machine learning model. Changing from one optimization algorithm to another, would not change the purpose of the training of the machine learning model. Therefore, this argument is unpersuasive and the grounds of rejection is herein maintained, albeit updated to reflect Applicant’s amendments to the claims.
Applicant, in the “Reservation of Rights” and “Conclusion” sections, does not present substantial arguments in support of the patentability of the claims. Therefore, said dependent claims stand rejected under the grounds of rejection presented herein and no detailed rebuttal is required.
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 23-27 are rejected under 35 U.S.C. § 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Step 1: Claims 23-27 are directed towards a method. Thus, these claims, on their face, are directed to one of the statutory categories of 35 U.S.C. § 101.
Step 2A - Prong One: As per MPEP 2106.04, Prong One asks does the claim recite an abstract idea, law of nature, or natural phenomenon. In Prong One examiners evaluate whether the claim(s) recites a judicial exception; that is, whether the claim(s) set forth or describe a law of nature, natural phenomenon, or abstract idea.
Claim 23 is presented here as a representative claim for specific analysis (The underlined claim terms here are interpreted as additional elements beyond the abstract idea.):
A method for training a machine learning model, comprising:
receiving, at a server, first data comprising a first plurality of features including at least one feature characterizing a user and at least an additional feature characterizing a client device of the user;
generating a feature vector based on an encoding of the first plurality of features;
using a mini-batch limited-memory optimizer, training the machine learning model in a first stage on the first data to characterize a likelihood of the user interacting with a content presentation displayed to the user during execution of a mobile game;
periodically receiving, at the server, incremental data comprising a second plurality of features including at least one feature characterizing an in-app event generated during the execution of the mobile game;
and using the mini-batch limited-memory optimizer, updating the machine learning model in a second stage with the incremental data, response to periodically receiving the incremental data.
The claims here are based on the recitation of an abstract idea (i.e. recitation other than the additional elements delineated here with underlining and further addressed per Step 2A - Prong Two and Step 2B). The claims recite the abstract idea of characterizing a likelihood of a user interacting with a content presentation which falls within certain methods of organizing human activity.
The phrase "certain methods of organizing human activity" applies to fundamental economic principles or practices including hedging insurance, mitigating risk; commercial or legal interactions including agreements in the form of contracts, legal obligations, advertising, marketing or sales activities or behaviors business relations; managing personal behavior or relationships or interactions between people including social activities teaching, and following rules or instructions. Refer to MPEP 2106.04(a)(2) II. A-C.
The Remaining Claims: The dependent claims recite fail to recite any additional elements beyond those already identified. The dependent claims further reiterate the same abstract idea with further embellishments: details the features in the feature vector (claim 24); in-app event is an in-app purchase (claim 25); types of games envisioned (claim 26); a more specific optimizer (claim 27). Therefore, the identified claims fall within the subject matter groupings of abstract ideas enumerated in MPEP 2106.04(a)(2).
Step 2A - Prong Two: As per MPEP 2106.04.II.A.2, Prong Two determines if the claim(s) recite additional elements that integrate the judicial exception into a practical application.
As for the additional elements of: a server, client device. To be patent-eligible, the elements additional to the identified abstract idea must amount to more than "an instruction to apply the abstract idea . . . using some unspecified, generic computer" to render the claim patent-eligible. Alice Corp. v. CLS Bank Int'l, 573 U.S. 208, 226 (2014). It would have been readily apparent to one having ordinary skill in the art (PHOSITA) at the time the invention was filed that the additional elements represent generic computing devices. Therefore, the claims amount to no more than a mere method, system, and/or computer program product to implement the abstract idea on a generic computer system. See MPEP § 2106.05(f).
As for the additional element(s) of: a mini-batch limited-memory optimizer, a machine learning model. The use of machine learning or artificial intelligence, without providing details of how the models themselves are improved, represents mere instructions to apply an exception. See MPEP 2106.05(f).
As for the additional element(s) of: receiving…first data, periodically receiving…incremental data. The gathering of data represents insignificant extra-solution activity that comprises mere data gathering. The gathering of data represents insignificant extra-solution activity that comprises mere data gathering. The additional element(s) represent insignificant extra-solution activity incidental to the primary process or product that are merely a nominal or tangential addition to the claim as noted in MPEP 2106.05(g).
As for the additional element(s) of: in-app, mobile game. These additional elements amounts to generally linking the use of the abstract idea to a particular technological environment or field of use (MPEP 2106.05(h)). The ordered combination offers nothing more than employing a generic configuration of computer devices and computer functions. The claims do not amount to a practical application, similar to how limiting the abstract idea in Flook to petrochemical and oil-refining industries was insufficient.
Step 2B: As per MPEP 2106.05, the additional elements are analyzed, both individually and in combination, to determine whether an "inventive concept" is furnished by an element or combination of elements that is recited in the claim in addition to (beyond) the judicial exception, and is sufficient to ensure that the claim, as a whole, amounts to significantly more than the judicial exception itself.
As for the additional element(s) of: receiving…first data, periodically receiving…incremental data. These elements represent content similar to receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information)
As for the additional element(s) of: a mini-batch limited-memory optimizer, a machine learning model. Applying "machine learning" at a high level of generality represents performing repetitive calculations, Flook, 437 U.S. at 594, 198 USPQ2d at 199 (recomputing or readjusting alarm limit values); Bancorp Services v. Sun Life, 687 F.3d 1266, 1278, 103 USPQ2d 1425, 1433 (Fed. Cir. 2012). Machine learning is well-understood, routine and conventional as exemplified in "Approaches to Machine Learning" by Langley et al. (Langley, P. and Carbonell, J.G. (1984), Approaches to machine learning. J. Am. Soc. Inf. Sci., 35: 306-316. https://doi.org/10.1002/asi.4630350509 (Year: 1984)). Furthermore, the Wikipedia article, “Limited-memory BFGS”, describes a “mini-batch limited-memory optimizer” technique that was extant prior the date the instant application was filed.
References of Record but not Applied in the Current Grounds of Rejection
The prior art listed below is made of record as considered pertinent to applicant's disclosure and is not relied upon in the grounds of rejection presented in this Office action. Those starred with '*' were added to this list in this Office action. Those without "*" were added in a previous Office action and are not repeated on a PTO-892 Notice of References Cited form, but are maintained herein for informational purposes only.
Jonathan Vaughan (Pub. #: WO 2022/256547 A1) discloses an advertisement selection system that uses "situational context of the user" to compute a "conversion likelihood probability" that is compared to a "threshold" (See 0052-055) and if the probability exceeds the threshold performs additional calculations including a "second layer of the algorithm" that includes "commercial significance" (See 0065).
Abraham Bagherjeiran, in "Method and System for Ranking Contextual Advertisements Based on Conversion Value", discloses an advertisement selection system that includes matching advertisements with prospective advertisement spaces in a web page by matching features of both the web page and the advertisement using conversion values based on advertiser's costs.
Examiner's Note on the Format of the Prior Art Rejections
The prior art rejections below contain underlined markings of the limitations (e.g. sample limitation). The underlined portions of a claim are addressed at the end of the grounds of rejection for that claim. Examiner notes that the underlining of the claim language is not a statement that the primary reference does not teach that language, but simply that said claim language is addressed at the end of the grounds of rejection for that claim.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
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) 23-25, 27 is/are rejected under 35 U.S.C. 103 as being unpatentable over:
Acuna Agost et al. (Pub. #: US 2019/0080260 A1) in view of
Jiang et al. (Pub. #: US 2023/0089895 A1) in view of
Narsky et al. (Pub. #: US 11,410,073 B1).
Claim(s) 23:
A method for training a machine learning model, comprising:
(Acuna Agost teaches displaying a webpage with advertisements to a device of a user in at least 0043 with request for such information made explicit in 0044-0046 with the disclosure of an ad exchange server receiving a request.)
receiving, at a server, first data comprising a first plurality of features including at least one feature characterizing a user
(Acuna Agost teaches that the request includes "user information relating to the user of the terminal device 126" in at least 0045. Also see 0059.)
and at least an additional feature characterizing a client device of the user;
(Acuna Agost teaches that the information regarding the user includes information such as "client-side information, such as device and browser identity and technical details" in at least 0045 and information regarding a "user device" in at least 0059.)
generating a feature vector based on an encoding of the first plurality of features;
(Acuna Agost teaches computing a feature vector in Figure 6, 0011, 0019, utilizing the features "from the raw values" in at least 0061. Also see 0062-0064.)
using a mini-batch limited-memory optimizer, training the machine learning model in a first state on the first data to characterize a likelihood of the user interacting with a content presentation displayed to the user during execution of a mobile game;
(Acuna Agost teaches that the output of the model using feature vectors is "an estimate of the likelihood of user interaction with an offer within a selected ad, based on the enriched feature vector" in at least 0078.)
periodically receiving, at the server, incremental data comprising a second plurality of features including at least one feature characterizing an in-app event generated during the execution of the mobile game;
(Acuna Agost teaches that the information regarding the user can include “feature values” that correspond to an “interaction event [that] occurred corresponding with the content placement event” in at least 0011. Also see 0011-0014)
and using the mini-batch limited-memory optimizer, updating the machine learning model with in a second stage incremental data, responsive to periodically receiving the incremental data.
(Acuna Agost discloses updating the learning model with new data in at least 0022, 0052, 0053, 0077, and claims 5, 10, and 15. Also see Figure 4.)
As for, "during execution of a mobile game", and “an in-app event generated during the execution of the mobile game;”: Acuna Agost discloses advertising "in-app" in at least 0002 and within a "mobile app" in at least 0010. Acuna Agost does not appear to make explicit that the mobile app is a "mobile game". However, Jiang teaches computing a user response prediction for displaying ads, including tracking events related to those ads, and that the ads can be within a "gaming application" in at least 0026 and 0039-0043.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the advertisement selection for display within a "mobile app" of Acuna Agost with the selection of advertisements within a "gaming application" as taught by Jiang. Motivation to combine Acuna Agost with Jiang comes from the desire to improve the "efficiency of automatic bidding platforms" (Jiang: 0003-0005).
As for, "using a mini-batch limited-memory optimizer", and "using the mini-batch limited-memory optimizer": Acuna Agost does not appear to specify a mini-batch limited-memory optimizer. However, Narsky teaches a technique of using an optimization function that is a "limited memory Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm" in at least Col. 5, Ll. 8-53. Examiner notes that the optimizer is described as: "(e.g., using a suitable optimizer, such as, for example, mini-batch Limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) or the like)" from paragraph 0023 of the specification.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the advertisement selection based on machine learning algorithms of Acuna Agost with the specific algorithm taught by Narsky. Motivation to combine Acuna Agost with Narsky comes from the desire to reduce the "utilization of computing resources" (Narsky: Col. 2, Ll. 16-20).
Claim(s) 24:
wherein the first plurality of features comprises a plurality of categorical features extracted from the first data, the plurality of categorical features including a publisher application, a device identifier, a timestamp, and a geographical location.
(Acuna Agost discloses event parameters that includes “publisher identifier” and “advertiser identifier” which corresponds to ‘publisher application’; “user device” which corresponds to a ‘device identifier’; “a user time segment” which corresponds to a ‘timestamp’; and “location” which corresponds to ‘a geographical location’ in at least 0058-0060 and Table 1.)
Claim(s) 25:
wherein the in-app even comprises an in-app purchase.
Acuna Agost does not appear to specify an event that comprises an in-app purchase. However, Jiang teaches a technique of tracking user responses such as “making a purchase” in order to improve advertising effectiveness in at least 0026, 0039, and 0043.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the advertisement selection for display within a "mobile app" of Acuna Agost with the selection of advertisements within a "gaming application" including the technique of tracking user purchases in-app as taught by Jiang. Motivation to combine Acuna Agost with Jiang comes from the desire to improve the "efficiency of automatic bidding platforms" (Jiang: 0003-0005).
Claim(s) 27:
wherein the mini-batch limited-memory optimizer comprises a mini-batch Limited-memory Broyden-Fletcher-Goldfarb-Shanno (L_BFGS) optimizer.
As for, “wherein the mini-batch limited-memory optimizer comprises a mini-batch Limited-memory Broyden-Fletcher-Goldfarb-Shanno (L_BFGS) optimizer”: Acuna Agost does not appear to specify a mini-batch limited-memory optimizer. However, Narsky teaches a technique of using an optimization function that is a "limited memory Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm" in at least Col. 5, Ll. 8-53.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the advertisement selection based on machine learning algorithms of Acuna Agost with the specific algorithm taught by Narsky. Motivation to combine Acuna Agost with Narsky comes from the desire to reduce the "utilization of computing resources" (Narsky: Col. 2, Ll. 16-20).
Claim(s) 26 is/are rejected under 35 U.S.C. 103 as being unpatentable over:
Acuna Agost et al. (Pub. #: US 2019/0080260 A1) in view of
Jiang et al. (Pub. #: US 2023/0089895 A1) in view of
Narsky et al. (Pub. #: US 11,410,073 B1) in view of
Burgin et al. (Pub. #: US 2014/0316870 A1).
Claim(s) 26:
wherein the mobile game comprises an asynchronous or synchronous competitive skill-based game.
Acuna Agost discloses advertising "in-app" in at least 0002 and within a "mobile app" in at least 0010. Jiang teaches that the mobile app can be a "mobile game". Acuna Agost, in view of Jiang, does not appear to specify that the game is “an asynchronous or synchronous competitive skill-based game”. However, Burgin discloses a technique of conducting a competitive skill-based game that is asynchronous or synchronous in nature in at least 0042-0047.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the advertisement selection based on machine learning algorithms in a game application of Acuna Agost, in view of Jiang and Narsky, with gaming features taught by Burgin. Motivation to combine Acuna Agost, in view of Jiang and Narsky, with Burgin derives from the desire to increase user engagement with advertisements (Burgin: 0004).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SCOTT SNIDER whose telephone number is (571)272-9604. The examiner can normally be reached M-W: 9:00-4:30 Mountain (11:00-6:30 Eastern).
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, Waseem Ashraf can be reached at (571)270-3948. 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.
/SCOTT SNIDER/Examiner, Art Unit 3621