DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . 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.
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 of this title, 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 1-3, 6-8, 12-15, 18-21, 24-27, and 30-31 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnamurthy et al. (US Pub. No. 2017/0282063), hereinafter referred to as Krishnamurthy, in view of Nice et al. (US Pub. No. 2014/0181121), hereinafter referred to as Nice.
Referring to claims 1, 7, 13, and 19, Krishnamurthy discloses one or more processors ([0073]), comprising: circuitry to: generate a plurality of feature vectors using game data associated with gameplay for one or more players of a game, wherein the plurality of feature vectors are generated over a time window of gameplay; the plurality of feature vectors representing one or more cumulative changes of a state of the game over the time window (monitors interactions of players with video games over time, [0023]; game session data 114 includes data about how the player 110 is playing a video game during a game session…time information about the game session, [0026]; the training dataset is composed of feature vectors. Each vector captures a set of features per player for a particular point of a video game (e.g., a point in time or a video game frame), [0070]; inputting the video game session data as input feature vectors to the neural network, [0085]); predict, using one or more neural networks, a next state for the game (predicted outcomes are derived based on automated analysis of a database of historical interactions, [0006]; artificial intelligence model 510 is trained to predict the potential outcomes, [0052]; fig. 6, [0058]); and generate one or more recommendations to the one or more players based, at least in part, on the next state (the computer system selects an action for the assistance from potential actions… computer system causes a presentation about the selected action in the video game in response to detecting that the video game player needs the assistance, [0006]; Based on these likelihoods, the action with the maximum likelihood or a number of top likelihoods (e.g., the top three) can be recommended to the player; [0062]).
Krishnamurthy does not appear to encode the plurality of feature vectors into a latent space; predict a next state feature vector using the latent space; and generate one or more recommendations based, at least in part, on the next state feature vector.
However, in a similar endeavor of making recommendation to users/players, Nice discloses encode the plurality of feature vectors into a latent space (latent space models that mathematically generate additional realized relationships between users and items…models a latent space using feature vectors of users and items, [0013]) predict a feature vector using the latent space (an item-latent-trait vector based on the item-stem vector and an item-offset vector is generated, [0046]), and generate one or more recommendations based, at least in part, on the feature vector (one or more recommended-media content identified based on the item-latent-trait vector are provided, [0046]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Krishnamurthy and Nice before him or her, to modify the recommendation system of Krishnamurthy to include the latent space modeling of Nice because the latent space would provide understanding of derived relationships.
The suggestion/motivation for doing so would have been to enable understanding of derived relationships (Nice: [0038]).
Therefore, it would have been obvious to combine Krishnamurthy and Nice to obtain the invention as specified in the instant claim.
As to claims 2, Krishnamurthy discloses the one or more circuits are further to identify one or more relationships (training dataset is composed of feature vectors; [0070]) between two or more events in the game that represent the one or more of cumulative changes of the state of the game (Each vector captures a set of features per player for a particular point of a video game (e.g., a point in time or a video game frame)…a feature vector includes interactions of the player with contexts, the contexts, outcomes of the interactions…an applicable feature vector includes actions performed in the video game in response to the player's interactions, the interactions, contexts associated with the actions, outcomes of the actions; [0070]).
As to claims 3, 15, 21, and 27, Krishnamurthy discloses the one or more of cumulative changes of the state of the game corresponds to at least one of game event data, gameplay data, statistical data, the training dataset is composed of feature vectors. Each vector captures a set of features per player for a particular point of a video game (e.g., a point in time or a video game frame)…a feature vector includes speed and types of actions performed, progress in a video game, accomplishments achieved, number of attempts per accomplishment, scores, player profile...a feature vector includes interactions of the player with contexts, the contexts, outcomes of the interactions, the experience level...feature vector includes actions performed in the video game in response to the player's interactions, the interactions, contexts associated with the actions, outcomes of the actions, the experience level...a feature vector includes type, presentation format, timing, an interactivity level of a context, the experience level; [0070]).
As to claims 6, 12, 18, 24, and 30, Krishnamurthy discloses the circuitry is further to provide the one or more recommendations for presentation to the one or more players (Based on these likelihoods, the action with the maximum likelihood or a number of top likelihoods (e.g., the top three) can be recommended to the player; [0062]).
As to claims 8, 14, 20, and 26, Krishnamurthy discloses the one or more circuits are further to receive data for one or more input types and transform the data into a plurality of feature vectors corresponding to a common schema (training dataset is composed of feature vectors. Each vector captures a set of features per player for a particular point of a video game...feature vector can also vary depending on the specific video game or genre of the video game; [0070]).
Referring to claim 25, Krishnamurthy discloses a player coaching system, comprising: one or more processors ([0073]) to: generate a plurality of feature vectors using game data associated with gameplay for one or more players of a game, wherein the plurality of feature vectors are generated over a time window of gameplay; the plurality of feature vectors representing one or more cumulative changes of a state of the game over the time window (monitors interactions of players with video games over time, [0023]; game session data 114 includes data about how the player 110 is playing a video game during a game session…time information about the game session, [0026]; the training dataset is composed of feature vectors. Each vector captures a set of features per player for a particular point of a video game (e.g., a point in time or a video game frame), [0070]; inputting the video game session data as input feature vectors to the neural network, [0085]); predict, using one or more neural networks, a next state for the game (predicted outcomes are derived based on automated analysis of a database of historical interactions, [0006]; artificial intelligence model 510 is trained to predict the potential outcomes, [0052]; fig. 6, [0058]); and generate one or more recommendations to the one or more players based, at least in part, on the next state (the computer system selects an action for the assistance from potential actions… computer system causes a presentation about the selected action in the video game in response to detecting that the video game player needs the assistance, [0006]; Based on these likelihoods, the action with the maximum likelihood or a number of top likelihoods (e.g., the top three) can be recommended to the player; [0062]); and memory for storing network parameters for the one or more neural networks (artificial intelligence model utilizes a number of neural networks, [0022]; values of such parameters are captured in a game session data...game session data is input to an artificial intelligence model...this model can be stored locally at the video game console; [0039]).
Krishnamurthy does not appear to encode the plurality of feature vectors into a latent space; predict a next state feature vector using the latent space; and generate one or more recommendations based, at least in part, on the next state feature vector.
However, in a similar endeavor of making recommendation to users/players, Nice discloses encode the plurality of feature vectors into a latent space (latent space models that mathematically generate additional realized relationships between users and items…models a latent space using feature vectors of users and items, [0013]) predict a feature vector using the latent space (an item-latent-trait vector based on the item-stem vector and an item-offset vector is generated, [0046]), and generate one or more recommendations based, at least in part, on the feature vector (one or more recommended-media content identified based on the item-latent-trait vector are provided, [0046]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Krishnamurthy and Nice before him or her, to modify the recommendation system of Krishnamurthy to include the latent space modeling of Nice because the latent space would provide understanding of derived relationships.
The suggestion/motivation for doing so would have been to enable understanding of derived relationships (Nice: [0038]).
Therefore, it would have been obvious to combine Krishnamurthy and Nice to obtain the invention as specified in the instant claim.
As to claim 31, Krishnamurthy discloses the circuitry is to generate the one or more recommendations (Based on these likelihoods, the action with the maximum likelihood or a number of top likelihoods (e.g., the top three) can be recommended to the player; [0062]) based, at least in part, on identifying a relationship between a current state of the game and a previous state of the game (the computer system predicts and selects parameters of the virtual assistant based on an assistant generator that implements a neural network. In this illustration, a portion of the player's features is input to the neural network as a feature vector. These features include…the current state of the video game…and previous sequence of interactions of the player, [0077]).
While Krishnamurthy teaches using feature vector representing the one or more of cumulative changes of the state of the game, Krishnamurthy does not appear to explicitly disclose the feature vectors are encoded into the latent space.
However, in a similar endeavor of making recommendation to users/players, Nice discloses encoding feature vectors into a latent space ([0013]).
The suggestion/motivation remains as indicated above.
Claims 5, 11, 17, 23, and 29 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnamurthy in view of Nice, as applied to claims 1-4, 6-8, 10, 12-16, 18-22, 24-28, and 30-31 above, further in view of Spies et al. (US Pub. No. 2021/0342496), hereinafter referred to as Spies.
As to claims 5, 11, 17, 23, and 29, while Krishnamurthy disclose the one or more neural networks accept input and generate the one or more recommendations based at least in part upon the one or more of cumulative changes of the state of the game, and Nice teaches a latent space, the combination of Krishnamurthy in view of Nice does not appear to explicitly disclose a generative adversarial network (GAN) to accept the latent space as input.
However, in a similar endeavor of making recommendations to users/players, Spies teaches a generative adversarial network (GAN) to accept the latent space as input (generator network 106 may generate an output image 114 based on...the latent space vector...the generator network 106 is a neural network...generator network 106 may be included in a generative adversarial network, [0024]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Krishnamurthy, Nice, and Spies before him or her, to modify the recommendation system of Krishnamurthy in view of Nice to include the GAN of Spies in order to generate results to guide a user.
The suggestion/motivation for doing so would have been to generate fine-tuned results (Spies: [0033])
Therefore, it would have been obvious to combine Krishnamurthy, Nice, and Spies to obtain the invention as specified in the instant claim.
Claims 4, 9-10, 16, 22, and 28 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnamurthy in view of Nice, as applied to claims 1-3, 6-8, 12-15, 18-21, 24-27, and 30-31 above, further in view of Misu et al. (US Pub. No. 2020/0039521), hereinafter referred to as Misu.
As to claims 4, 10, 16, 22, and 28, while Krishnamurthy teaches a time window used to predict the next state (monitors interactions of players with video games over time, [0023]; game session data…time information about the game session, [0026]; inputting the video game session data as input feature vectors to the neural network, [0085]; artificial intelligence model 510 is trained to predict the potential outcomes, [0052]), and Nice teaches considering time-based behavior ([0027-0028], [0042]) and predicting a feature vector using the latent space (an item-latent-trait vector based on the item-stem vector and an item-offset vector is generated, [0046]), the combination of Krishnamurthy and Nice does not appear to explicitly disclose determining a size for a time window.
However, in a similar endeavor of evaluating behavioral events, Misu discloses determining a size for a time window to be used (evaluate a plurality of various time window sizes…determine one or more optimum context time window sizes to be utilized, [0074]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Krishnamurthy, Nice, and Misu before him or her, to modify the recommendation system of Krishnamurthy in view of Nice to include window size evaluation of Misu in order to analyze the relationship between the context window size and prediction performance.
The suggestion/motivation for doing so would have been to determine the optimum time window size (Misu: [0074])
Therefore, it would have been obvious to combine Krishnamurthy, Nice, and Misu to obtain the invention as specified in the instant claim.
As to claim 9, Krishnamurthy discloses the one or more of cumulative changes of the state of the game are based, at least in part, on relationships between two or more events within the game (a feature vector includes speed and types of actions performed, progress in a video game, accomplishments achieved, number of attempts per accomplishment, scores, player profile...a feature vector includes interactions of the player with contexts, the contexts, outcomes of the interactions, the experience level...feature vector includes actions performed in the video game in response to the player's interactions, the interactions, contexts associated with the actions, outcomes of the actions, the experience level...a feature vector includes type, presentation format, timing, an interactivity level of a context, the experience level; [0070]; a portion of the player's features is input to the neural network as a feature vector. These features include the player's experience level, the current state of the video game (e.g., a video game frame, an identifier of a context, etc.), and previous sequence of interactions of the player, [0077]).
While Krishnamurthy discloses the two or more events each occur within a corresponding event window (Each vector captures a set of features per player for a particular point of a video game; [0070]), Krishnamurthy does not appear to explicitly disclose a window size calculated using the one or more neural networks.
However, in a similar endeavor of evaluating behavioral events, Misu discloses a window size calculated using the one or more neural networks (evaluate a plurality of various time window sizes…determine one or more optimum context time window sizes to be utilized, [0074]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of Krishnamurthy, Nice, and Misu before him or her, to modify the recommendation system of Krishnamurthy in view of Nice to include window size evaluation of Misu in order to analyze the relationship between the context window size and prediction performance.
The suggestion/motivation for doing so would have been to determine the optimum time window size (Misu: [0074])
Therefore, it would have been obvious to combine Krishnamurthy, Nice, and Misu to obtain the invention as specified in the instant claim.
Response to Arguments
Applicant's arguments filed 3/16/2026 have been fully considered but they are not persuasive.
With respect to claim 1, rejected under 35 U.S.C. 103 as unpatentable over Krishnamurthy in view of Nice, and regarding the feature vectors taught by Krishnamurthy, on pg. 11 of the response the Applicant asserts:
“…These feature vectors seem only to capture current states of a game and not a "cumulative changes of a state of the game over the time window," much less state changes that have been encoded to a latent space, as recited in claim 1. Nothing in Krishnamurthy seems to contemplate changes of state in a game….”
The Examiner respectfully disagrees. As demonstrated in the rejections above, Krishnamurthy teaches monitoring the players interaction with video games over time, collection game session data, and using the video game session data as input feature vectors.
Regarding the teachings of Krishnamurthy in view of Nice, on pg. 11 of the response the Applicant asserts:
“…While Nice mentions that user to content relationships are embedded to model derived relationships among users, items, and metadata, Nice does not describe accumulated game-state changes, gameplay state prediction, or generating a future game state for a player of a game, from latent-space gameplay data. Therefore, Nice does not teach, "encod[ing] the plurality of feature vectors into a latent space representing one or more cumulative changes of a state of the game over the time window," or "predict[ing], using one or more neural networks, a next state feature vector for the game using the latent space," as recited by claim 1.”
In response to applicant's arguments against the references individually (i.e., “Nice does not describe accumulated game-state changes, gameplay state prediction, or generating a future game state for a player of a game”), one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). As demonstrated in the rejections above, Krishnamurthy was presented as teaching the gameplay elements, while Nice was relied upon to show a teaching of latent space modeling features, and the limitations of the claims were considered obvious based on the combination of the references.
The Applicant further asserts:
“Further, the combination of references is improper. The rejection combines the references alleging that Nice's latent-space techniques could be incorporated into the recommendation system of Krishnamurthy to better capture derived relationships. Applicant respectfully disagrees. Krishnamurthy's feature vectors represent game information at particular points in time, while Nice's latent space is not directed to accumulated game-state changes or future game-state prediction in a gameplay setting, but instead simply an embedding of relationships between users and content. As a result, the combination does not clearly show the claimed architecture in which a plurality of accumulated state changes in the game are encoded to a latent space and a predicted future state is generated by the neural network using that latent space.”
Again, in response to applicant's arguments against the references individually (i.e., “Nice's latent space is not directed to accumulated game-state changes or future game-state prediction in a gameplay setting”), one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). As demonstrated in the rejections above, Krishnamurthy was presented as teaching the gameplay elements, while Nice was relied upon to show a teaching of latent space modeling features, and the limitations of the claims were considered obvious based on the combination of the references. The Applicant’s assertion that “the combination of references is improper” and “the combination does not clearly show the claimed architecture” amount to general allegations that are not substantiated by the arguments against the references individually.
The remaining remarks rely on the persuasiveness of the remarks directed to the independent claims, and therefore are not persuasive as well.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. The US Pub. No. 2020/0076841 of Hajimirsadeghi et al. and the US Pub. No. 2021/0390316 of Schoneveld et al. are pertinent to generating predictive vectors.
The examiner has cited particular column, line, and/or paragraph numbers in the references as applied to the claims above for the convenience of the applicant. Although the specified citations are representative of the teachings of the art and are applied to the specific limitations within the individual claim, other passages and figures may apply as well. It is respectfully requested from the applicant in preparing responses, to fully consider the references in its entirety as potentially teaching of all or part of the claimed invention, as well as the context of the passage as taught by the prior art or disclosed by the examiner.
The examiner requests, in response to this office action, support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line number(s) in the specification and/or drawing figure(s). This will assist the examiner in prosecuting the application. When responding to this office action, applicant is advised to clearly point out the patentable novelty which he or she thinks the claims present, in view of the state of art disclosed by the references cited or the objections made. He or she must also show how the amendments avoid such references or objections. See 37 C.F.R. 1.111(c).
Applicants seeking an interview with the examiner, including WebEx Video Conferencing, are encouraged to fill out the online Automated Interview Request (AIR) form (http://www.uspto.gov/patent/uspto-automated-interview-request-air-form.html). See MPEP §502.03, §713.01(11) and Interview Practice for additional details.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ERIC T OBERLY whose telephone number is (571)272-6991. The examiner can normally be reached on M-F 800am-430pm (MT).
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Dr. Henry Tsai can be reached on (571) 272-4176. 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.
/ERIC T OBERLY/ Primary Examiner, Art Unit 2184