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 Amendment
Applicant has amended the claims and argues Thomas does not generalize new situations (Remarks 6 filed 9/11/2026). Examiner agrees and presents for Applicant’s consideration Aghdaie in view of Crowley.
Applicant has not provided arguments for the rejections under 35 USC § 101, thus the rejection of the claims are maintained.
Applicant has amended the claims thus the objection to Claims 1 and 14 are withdrawn.
Claim Rejections - 35 USC § 101
5. 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.
6. Claims 1-6, 8-10, 13-14, 16-18, and 20-25 are rejected under 35 U.S.C. § 101 because the claimed invention is directed to an abstract idea without significantly more.
7. Step 1
Claims 1-19 are directed to methods meeting the requirements for Step 1.
8. Step 2A Prong 1
Independent Claim 1 (below) recites “training a machine learning model using the gameplay data” among other details of the abstract idea Independent Claim 14 recites a similar limitation. This limitation recites a mathematical calculation, which is a mathematical concept, and thus, an abstract idea.
A method, comprising:
recording gameplay data from one or more sessions of a video game, the one or more sessions defined for interactive gameplay of a user, wherein the gameplay data comprises visual data and user inputs;
training a machine learning model using the gameplay data, wherein the machine learning model is an artificial neural network, wherein the training comprises a supervised learning task that configures the machine learning model to imitate the interactive gameplay of the user on a decision-by-decision basis;
after the training, exposing the machine learning model to one or more scenarios of the video game that are not defined from the recorded one or more sessions, such that the machine learning model generates gameplay decisions in response to the one or more scenarios;
evaluating the gameplay decisions of the machine learning model in response to the one or more scenarios to determine one or more descriptive features of the user's gameplay; and
using the determined descriptive features of the user's gameplay to provide a recommendation to the user.
9. Step 2A Prong II
There is no discernable additional element (or combination of elements) recited in Claims 1 or 14 that have integrated the judicial exception into a practical application.
In making this determination Examiner relies on the two step guidance from “Advance notice of change to the MPEP in light of Ex Parte Desjardins (December 5, 2025) wherein the claims were directed to a machine learning model on a series of tasks. First the specification must be evaluated to decide if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement “in the functioning of a computer, or an improvement to other technology or a technical field.” (Emphasis in original). Second, if the specification sets forth an improvement, the claim must be evaluated to ensure the claim itself reflects the disclosed improvement i.e., the components or steps of the invention that provide the improvement described in the specification.
In Desjardins, the specification explained how the machine learning model is trained to learn new tasks while protecting knowledge about previous tasks to overcome the problem of “catastrophic forgetting,” and that the claims reflected the improvement identified in the specification. Moreover, the specification disclosed enumerated improvements, namely: the effective learning of new tasks in succession in connection with specifically protecting knowledge concerning previously accomplished tasks; allowing the system to reduce use of storage capacity; and the enablement of reduced complexity in the system. “Such improvements were tantamount to how the machine learning model itself would function in operation and therefore not subsumed in the identified mathematical calculation” according to the guidance.
Here, examination of the specification indicates no such improvements in the functioning of a computer, or of an improvement to other technology or a technical field. According to Applicant in his specification the machine learning model is operating in a highly conventional computing environment. “It will be appreciated that the hardware utilized for the video game experience can vary in various implementations. For example, the computing device can be a game console, personal computer, laptop, tablet, cell phone, or any other type of computing device that is capable of executing a video game as described herein.” ([0038]). And that the learning model is recited as a tool to be applied where “In other implementations, the machine learning model 112 can utilize other types of artificial intelligence or machine learning techniques (e.g. support vector machine, linear regression, logistic regression, naive Bayes, decision trees, k nearest neighbor, etc.) that can be configured to imitate the gameplay of a user, in accordance with implementations of the present disclosure.” ([0050]).
But instead of describing improvements to the model itself on par with the articulated improvements of Desjardins, the specification describes uses, namely: how an AI trained bot can be deployed to play a game on the user’s behalf ([0053-0054]); provide for new experiences in game play to play against other’s AI bots ([0055]); to cover for a human user when the human user is not available during group play ([0056]); and for a user to borrow another player’s AI bot to help the player overcome a difficult section of a game ([0058]). In contrast to Desjardins, the Federal Circuit held machine learning claims ineligible in Recentive Analytics, Inc. v. Fox Corp., No. 23-2437 (Fed. Cir. 2025).
In Recentive, the claims were directed towards providing parameters to a machine learning model to train the model to identify relationships between different event parameters and target features using historical data corresponding to one or more previous series of live events, and changing conditions, and, using machine learning, dynamically generate optimized maps and schedules. However, the Court found that the technology described in the specification and recited in the claims was conventional. Neither the specification nor the claims described and recited, respectively, how any such improvements in AI was accomplished i.e., by an articulated specific technological improvement to the underlying machine learning method not just claimed uses in new environments. Consequently, the Court held the claims failed to provide “significantly more” than the abstract idea of generating event schedules and network maps through the application of machine learning.
In weighing these considerations, Examiner deems that Applicant’s disclosure and claims unlike those of Desjardins and similar to those of Recentive because the specification describes a conventional computing environment and a litany of machine models that can be performing as designed therein. Also, the specification and the claims, while expressing the benefits of using machine learning “to provide a recommendation to the user” (See Claim 1), fail to express and claim any technological improvement in the operation of the machine learning itself e.g., an improvement to the algorithm; reduction in complexity; or a savings via efficient use of storage space as in Dejardins.
Examiner deems the balance of the claimed steps to record video game play data from one of more sessions; to train the model on the data; to expose the model to one or more scenarios of the video game; to evaluate the game play for descriptive features; and to provide a recommendation to a user to be, in totality, a conventional recitation of the machine learning model and model training process applied to express news uses in the technological field of video gaming. Thus, these remaining claimed elements are deemed to be conventional machine model training and a technological field of use. Thus, Claims 1 and 14 lack a practical application.
10. Step 2B
According to the advanced guidance to be reflected in MPEP 2106, in addition to the considerations discussed in Step 2A, an additional consideration indicative of an inventive concept (aka “significantly more”) is the addition of a specific limitation other than what is well-understood, routine, conventional activity in the field (MPEP 2106.05(d)). Conversely, an additional consideration not indicative of an inventive concept is simply appending well-understood, conventional activities previously known to the industry, specified at a high level of generality, to the abstract idea (MPEP 2106.05(d) and Berkheimer Memo, April 20, 2018). Thus, the additional elements evaluated under Step 2A are re-evaluated in Step 2B to determine if they are more than what is well-understood, routine, conventional activity in the field. Examiner notes there are no extra-solution elements but merely limitations of a highly generic video game environment field of use. Thus, Claims 1 and 14 lack an innovative concept and are ineligible.
11. Dependent Claims 2-6, 8-10, 13, 16-18, and 20-25
Claims 2-6, 8-9, 13, 16-18, and 20-25 are features, skills, techniques, tendencies, opponents, styles, paths, skill levels, subsequent session recommendations, data types: video, video inputs, game states, and image frames related to either the inputs, outputs, or characterizations of the trained model. Claim 10 recites a well-known game controller device.
Claim Rejections - 35 USC § 103
12. 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.
13. 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.
14. The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied 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.
15. Claims 1-6, 8-10, 13-14, 17-18, and 20-25 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pat. Pub. No. 2018/0243656 to Aghdaie in view of U.S. Pat. Pub. No. 2009/0210078 to Crowley and U.S. Pat. No. 2018/0367484 to Rodriguez.
In Reference to Claims 1 and 14, 17, and 20
Aghdaie discloses a method (Fig.1, [0024]), comprising:
recording gameplay data from one or more sessions of a video game, the one or more sessions defined for interactive gameplay of a user, wherein the gameplay data comprises visual data and user inputs ([0171], see also state variables [0035, 0079, 0086], user actions performed in the course of gameplay [0127], “game state information 916, which can include character states, environment states, scene object storage, and/or other information associated with a state of execution of the game application 1010. For example, the game state information 916 can identify the state of the game application at a specific point in time, such as a character position, character action, game level attributes, and other information contributing to a state of the game application.” [0173]);
training a machine learning model (Fig. 2) using the gameplay data, wherein the machine learning model is an artificial neural network ([0078-0079, neural [0052]), wherein the training comprises a supervised learning task that configures the machine learning model to imitate the interactive gameplay of the user on a decision-by-decision basis (0052, 0056]);
after the training, exposing the machine learning model to one or more scenarios of the video game that are not defined from the recorded one or more sessions, such that the machine learning model generates gameplay decisions in response to the one or more scenarios (Fig. 2 206 prediction models resulting in difficulty trajectories Fig. 4B where “By tracking the trajectory of difficult for users who have progressed through the game, difficulty trajectories may be determined for other users that may not have progressed as far into the game.” [0133]);
evaluating the gameplay decisions of the machine learning model in response to the one or more scenarios to determine one or more descriptive features of the user's gameplay (Fig. 5 player progression, see also Fig. 6 608-610 monitoring progress and adjusting difficulty); and
using the determined descriptive features of the user's gameplay to provide a recommendation to the game designers (Fig. 9).
Aghdaie discloses the invention substantially as claimed. However, the reference does not explicitly disclose providing recommendations to users. One of skill in the art would be aware of the electronic analysis and athletic performance of Crowley.
Crowley teaches of athletic performance analysis to determine a skill level of an athlete with respect to one of more actions (Abstr., [0019]) to include using machine learning to identify correlations in data [0086, 0088] and providing recommendations to the user where “[f]or example, certain numerical figures or graphs may be generated to visually show where the athlete scores on a continuum of skill levels. In addition, recommendations may be generated in a textual format to be given to the athlete, with particular instructions on how the athlete can improve their performance, including suggested exercises or drills to perform in order to improve the athlete's muscle memory for a particular task. For example, if the testing of an athlete indicates that the athlete does not release the ball during a dribble with adequate velocity, the system 100 can recommend drills and conditioning routines to address such a situation.” [0044].
Crowley provides this invention to provide feedback for an athlete so that they can improve their performance ([0007]).
The Supreme Court in KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 415-421, 82 USPQ2d 1385, 1395-97 (2007) identified a number of rationales to support a conclusion of obviousness
(A) Combining prior art elements according to known methods to yield predictable results;
(B) Simple substitution of one known element for another to obtain predictable results;
(C) Use of known technique to improve similar devices (methods, or products) in the same way; and
(D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results.
Here, one of skill in the art would look to Crowley to modify the output of Aghdaie to utilize the information based on the evaluations to inform not only developers but also users to help improve their performance. The Courts have held that combining prior art elements according to known methods to yield predictable results to be indicia of obviousness.
If Applicant disagrees that the prediction trajectories of Aghdaie do not imitate interactive gameplay on a decision-by-decision basis, one of skill in the art would recognize the modeling teachings of Rodriguez.
Rodriquez teaches of game applications and video data [0008], game data and bot data [0053], and the use of machine learning to train machine learning models [0053, 0054]. “After the machine learning model is trained, a newly-occurring set of data can be input to the model and the model can provide suggested items based on its training with the sample data. Based on the sample data, the machine-learning model can predict messages and commands to occurring events in an embedded session, which may then be provided as suggested response items. User interaction is enhanced, e.g., by reducing burden on the user to determine a command or compose a message to an application event, by providing a choice of response items that are customized based on the occurring event and the user's context.” [0243].
It would require only routine skill in the art to amplify that the predicted trajectories of Aghdaie which are a culmination of the actions of the user and the game application for each scene for every occurrence as explained by Rodriguez to achieve the predictable result of improving the understanding of the predictions and enhance user interaction by providing a choice of responses {recommendations}. The Courts have held that combining prior art elements according to known methods to yield predictable results to be indicia of obviousness.
In Reference to Claims 2 and 25
Aghdaie discloses wherein the descriptive features of the user's gameplay identify a style or tendency of the user during gameplay ([0030, 0079]).
In Reference to Claim 3
Aghdaie discloses identifying skill level with respect to playing the video game [0026]. Further, Crowley teaches of providing feedback and constructive criticism ([0010, 0044]).
In Reference to Claims 4 and 21
Crowley discloses wherein the recommendation identifies an opponent for the user to play against where a report generator 210 compiles basic data on particular skills and corresponding data for the athlete to match players with respect to a curtained skill and provide feedback to let the athlete know where they should focus their training.
In Reference to Claim 5
Crowley teaches wherein the recommendation identifies a style of play for the user to
develop as in a manner in which an opponent dribbles a basketball strong to the left and weak
to the right ([0092]).
In Reference to Claims 6 and 22
Crowley teaches wherein the recommendation identifies a path of development for the user where an athlete is tested in multiple time periods so as to assess the progress of the athlete and to generate a prediction of the near term and longer term expected progress ([0037]).
In Reference to Claims 8 and 23
Aghdaie discloses wherein training the machine learning model uses the visual data and user inputs to cause the machine learning model to respond to a given portion of the visual data by generating inputs similar to the user inputs that were generated in response to the given portion of the visual data of the first session where the user data repository 138 may further comprise data from one or more automated agents running defined routines (“bots”) (not shown). Each bot may be used to simulate gameplay by stepping through a plurality of gameplay states of the video game 112. In some embodiments, different bots may be associated with different rules or procedures for stepping through the gameplay states of the video game 112 of a particular portion of the video game 112. In some embodiments, bots may be configured to step through gameplay states of particular portions of the video game 112 associated certain gameplay parameter values, in order to evaluate a difficulty level associated with the gameplay parameter values [0089].
In Reference to Claims 9 and 24
Aghdaie discloses scene object information from user gameplay [0127, 0173] where the scene is comprised of frames where “[t]he full scene can [sic be] conveyed to a renderer, which can generate a new frame for display to the user. The process can be repeated for rendering each frame during execution of the game application.” [0174].
In Reference to Claim 10
Aghdaie discloses wherein the user inputs are defined from game controllers [0179].
In Reference to Claim 13
Aghdaie discloses wherein the gameplay data includes game state data from the first session of the video game (“user actions performed in the course of gameplay” [0127]).
In Reference to Claim 18
Crowley teaches of providing recommendations both quick and minimal and longer and more in-depth ([0071]) and of comparing data on skills from an earlier time to a current time. One of skill in the art would be aware of providing timely feedback, even during the current time to modify a next game of Aghdaie to provide any feedback. The Courts have held that the use of a known technique to improve similar devices (methods, or products) in the same way to be indicia of obviousness.
16. Claims 16 is rejected under 35 U.S.C. 103 as being unpatentable over Aghdaie, Crowley, Rodriguez further in view of U.S. Pat. Pub. No. 2006/0246973 to Thomas.
Aghdaie discloses the invention substantially as claimed. However, the reference does not explicitly disclose tendencies of a user defined by activity and non-activity of the user. One of skill in the art would be aware of the teachings of Thomas.
Thomas discloses a user profile can contain sub-profiles of events ([0028]) wherein for example, “[a]n event 685, as mentioned above, can be defined by a particular action taken place either by a user or by the computer program 110 itself. An action can be movement by an avatar 687, special moves invoked 689, stoppage of time 691, such as a timeout, etc.” [0028].
Here, it would require only routine skill in the art when considering the parameters and variable definitions of a prediction model and approach of Thomas to amplify the meaning of model parameters of Aghdaie to include in the collection of user data to includegameplay to address activity and non-activity of the user’s gameplay behavior. The Courts have held that combining prior art elements according to known methods to yield predictable results to be indicia of obviousness.
Conclusion
17. THIS ACTION IS MADE FINAL. 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.
18. Any inquiry concerning this communication or earlier communications from the examiner should be directed to Paul A. D’Agostino whose telephone number is (571) 270-1992.
19. 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.
20. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Kang Hu can be reached on (571) 270-1344. The fax phone number for the organization where this application or proceeding is assigned is 571-270-2992.
/PAUL A D'AGOSTINO/ Primary Examiner, Art Unit 3715
