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 .
Priority
Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d). The certified copy has been filed in parent Application No. JP2021-200110, filed on 12/09/2021.
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.
Claim(s) 1-2, 4, 6, and 9-10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Osman et al. (U.S. Patent No 20180001205) in view of Mousavi et al.
Regarding claim 1 and substantially similar limitations in claims 9 and 10, Osman et al., henceforth Osman, discloses “an information processing device (the automated artificial intelligence (AI) personal assistance system 129) comprising: a learning unit (paragraph 54, lines 1-6, “profiler engine 145 is able to collect and analyze user data when a corresponding user 115A is playing any gaming application”) configured to learn a situation by machine learning (see paragraph 42, lines 18-22, “the profiler engine 145 is able to learn (e.g., by applying deep learning or artificial intelligence techniques) how a particular gaming application should be played (generally, or within the context of a given level, sub-level, or given problem)”; see also paragraphs 70, lines 5-8 and 9-13, “the neural network 190 represents an example of an automated analysis tool for analyzing data sets used by the user game play profiler 145A to determine game play profiles […] an instance of the neural network 190 may be used to train the user game play profiler 145A to determine the game play profile […] which can be used to control the game play of the user 115S" and 74, lines 5-8 and 14-17, “the output nodes may identify problems, tasks, task types, difficult tasks, approaches to complete a specific task within a gaming application, […] provide appropriate responses to take regarding a presented task or task type according to a particular user's game play style, learn appropriate responses or approaches to take with respect to a particular task, etc.”), wherein the situation is learned at a time of a user operation on an application (see paragraph 61, lines 2-5, “for every response taken by the user 115S of client device 100 to address a task or subtask or action, the action state 121, action metrics 122, and action results 123 are collected by AI personal assistant 120”, paragraph 62, lines 1-3, “the action state 121 defines the action, sub-task, or task that is addressed within the context or gaming environment of the gaming application”, and paragraph 63, lines 1-4, “The collected data (e.g., action state 121, action metrics 122, and actions results 123) are delivered to profiler engine 145 for further analysis (e.g., generating a user response to tasks of certain type)”), the user operation is associated with a user (see paragraph 63, lines 1-4, “The collected data (e.g., action state 121, action metrics 122, and actions results 123) are delivered to profiler engine 145 for further analysis (e.g., generating a user response to tasks of certain type)”), […] and an assistance unit configured to execute, across at least one application, operation assistance processing to automatically operate the application based on a learning result of the learning unit, wherein the at least one application includes the application (see paragraph 68, lines 1-7 and 8-11, “the game play controller 171 is configured to control an instance of a gaming application based on the user game play style of a corresponding user game play profile for a corresponding user 115S. In particular, when automated game play control is requested by the user 115S, detection of a first task that is presented to the user 115S is effected in order to determine an appropriate response […] The first task that is presented and/or approached in the game play of the user 115S is automatically controlled based on a defined user first response”).
However, Osman is silent on “the situation includes a mental state of the user and a physical state of the user.”
Mousavi et al., henceforth Mousavi, teaches a brain-computer interface system for controlling a cursor on a computer screen by employing machine learning to interpret the user’s brain activity. Notably, Mousavi’s system can detect, from the user’s brain activity, whether the system has correctly or incorrectly interpreted the user’s intentions, and adjusts the output of the system accordingly. Thus, Mousavi’s system can be said to employ machine learning to learn a correlation between “a mental state of the user with respect to the input or the output [of the system]” and is capable of changing the output of the system in response to the user’s perception of an erroneous input (“change operation at a time of an erroneous input”). Mousavi also teaches using error-related brain activity to improve the performance of a machine learning classifier in the future, essentially allowing the classifier to “learn from its mistakes” (“A second approach, based on error-driven learning, attempts to limit the occurrence of a future error by updating the classifiers upon the detection of error-related brain activity or to discard the unsupervised adaptation temporarily if an error is detected”). Finally, Mousavi teaches that systems which interpret users’ mental activity based on their brain waves have been known in the art of video games for at least several years at the time of writing (e.g., “For instance, the authors in [28] showed that error signals can be detected from human electrocorticography (ECoG) in a continuous task comprising a video game”).
Finally, Osman discloses the use of a head-mounted display as a client device, which includes “Bio-sensors 442 […] to enable detection of physiological data from a user. In one embodiment, the bio-sensors 442 include one or more dry electrodes for detecting bio-electric signals of the user through the user's skin” (paragraph 129, lines 1-5). Dry electrodes, as would be known to one of ordinary skill in the art, are capable of detecting the electrical activity of a human brain when placed on the scalp, and the inclusion of such dry electrodes in a head-mounted display means they would already be in position for recording the user’s brain activity. One of ordinary skill in the art would also be aware that dry electrodes are also capable of detecting the muscle movements of a user by picking up on the electrical signals generated when muscles contract, meaning that Osman’s device would further be inherently capable of detecting the “physical activity” of the user, in conjunction with the cameras, gyroscopes, and accelerometers embedded in or used alongside the head-mounted display (see paragraphs 119-123).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the information processing device of Osman to employ the sensors already included in the device to record the brain activity, muscle activity, and motion data of the user so that the device could accept the mental and physical state of the user as inputs to the machine learning model, as taught by Mousavi, and they would have been motivated to do so to allow Osman’s AI assistant to learn to identify, based on the user’s brain activity, when it has made a mistake while performing autonomous control of the user’s gameplay, as taught by Mousavi, and learn from this feedback to reduce the occurrence of future mistakes, as taught by Mousavi, with reasonable expectation of success.
Regarding claim 2, Osman (in view of Mousavi) discloses “the information processing device according to claim 1, wherein the application is a game application (see paragraph 4, lines 1-3, “Embodiments of the present disclosure relate to systems and method for providing gaming control to a user playing a gaming application”), the situation at the time of the user operation further includes information related to the situation at the time of play by the user (the “action state”, “action metrics”, and “action results” described by Osman in paragraph 62), wherein the situation is indicated by a specific application different from the game application (see paragraph 37, “In another embodiment, AI personal assistant 120 may be configured remote to client device 100, and independent of the game cloud system 210”), the learning unit is further configured to learn a correlation between each element of elements of the information related to the situation at the time of the play and information related to the game application (see paragraph 63, lines 1-4, “The collected data (e.g., action state 121, action metrics 122, and actions results 123) are delivered to profiler engine 145 for further analysis (e.g., generating a user response to tasks of certain type)”), and the assistance unit is further configured to perform, based on a learning result of the learned correlation, the automatic operation (see paragraph 68, lines 1-7 and 8-11, “the game play controller 171 is configured to control an instance of a gaming application based on the user game play style of a corresponding user game play profile for a corresponding user 115S. In particular, when automated game play control is requested by the user 115S, detection of a first task that is presented to the user 115S is effected in order to determine an appropriate response […] The first task that is presented and/or approached in the game play of the user 115S is automatically controlled based on a defined user first response”).
Regarding claim 4, Osman (in view of Mousavi) discloses “the information processing device according to claim 2, wherein the assistance unit is further configured to set contents of the automatic operation based on a feature associated with the user at the time of the play of the game application, and the feature is indicated by the learning result of the learned correlation (learning and emulation of the behavior of a player in response to a particular in-game situation by the AI assistant is disclosed by Osman in paragraphs 70, lines 9-13, “an instance of the neural network 190 may be used to train the user game play profiler 145A to determine the game play profile […] which can be used to control the game play of the user 115S" and paragraph 74, lines 5 and 14-16, “the output nodes may […] provide appropriate responses to take regarding a presented task or task type according to a particular user's game play style”).”
Regarding claim 6, Osman (in view of Wakeford) discloses “the information processing device according to claim 4, wherein the assistance unit is further configured to enable the user operation to interrupt execution of the automatic operation corresponding to the set contents” (see paragraph 19, lines 14-16, “When the user wishes to return to active play, the user can stop the AI character, and allow the user to continue the game”).
Claim(s) 3 is/are rejected under 35 U.S.C. 103 as being unpatentable over Osman in view of Mousavi, as applied to claim 2 above, and further in view of Wakeford et al. (U.S. Patent No 10286322) and Elenbaas et al. (US Patent No. 20210129032).
Regarding claim 3, Osman (in view of Mousavi) discloses “the information processing device according to claim 2, wherein the elements include: a game title of the game application (Osman’s system includes a game title processing engine; see paragraph 45: “Game title processing engine 130B is able to support a plurality of gaming applications using a plurality of game logics 177”), an input by the user during the play of the game application (the “action metrics” described by Osman in paragraph 62), an output corresponding to the input, the mental state of the user with respect to one of the input or the output (the output of the gaming application in response to a user’s input would be included in the “responses taken by the user” associated with Osman’s “action metrics”).”
However, Osman does not disclose “the elements include: a game title of the game application, […] the physical state of the user with respect to one of the input or the output, position information of the user, a temperature around the user, and information indicating correction contents corresponding to an erroneous input.”
Wakeford et al., henceforth Wakeford, describes a system for learning the behaviors of a player of a mobile game application in response to in-game events so that the system may automatically execute certain game actions on behalf of the player in situations when the player is not available or is otherwise preoccupied. Wakeford teaches collecting data from external sources outside of the game application, such as sensors worn on the user’s body, to determine when the system should take action on behalf of the user (see column 12, lines 49-58: “[…] actions may be determined by action component 120 based on user activity information regarding a users' current activity in the real-world by the given user. The user activity information may be obtained by action component 120 from user component 116 and/or other sources. The user activity information, for example, may indicate that the user is performing one or more real-world activities (e.g., driving, sleeping, etc.) and/or whether or not the user may be available to interact with the online game via a client game application” and column 10, lines 44-48, “In some implementations, the user component 116 may obtain and/or receive the user activity information from personal devices associated with the users, such as client computing platform(s) 104, wearable device 124, sensors, and/or any other type of personal device associated with the given user”). Wakeford also teaches that the user activity information may include information regarding the location of the user (see column 8, lines 21-23, “[…] a real-world location where the user took the action may be monitored, tracked, and/or stored by gameplay component 114”).
Finally, Osman provides a motivation for a user to employ artificial intelligence to automate their own gameplay, namely allowing the AI assistant to temporarily “take over” for the user when the user wishes to take a break, “maintaining a presence” in the online game for the user while freeing them to attend other tasks external to the game (see paragraph 19, lines 21-29, “in a competitive multi-player game the user can allow the game play controller to take over game play of the user, so that the user is able to break away from the game (e.g., to eat, work, etc.). The game play controller trains an AI character to simulate the game play of the user (e.g., playing style), so that their online presence is maintained in the multi-player gaming session, and so that progress can continue in the gaming application”). It should be noted that this motivation aligns rather closely with the motivation provided by Wakeford, which also describes a user not being present in an online game as potentially being undesirable for the user (see column 1, lines 17-24, “if a user cannot or does not want to launch the game application, the user may not be able to respond and/or initiate actions in response to things that occur within the online game. As such, users that are not able to or do not want to access, launch, and/or run the game application may be disadvantaged as events and/or occurrences happen within the online game, and the user cannot respond”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the information processing device of Osman to use data collected from external sources outside of the game application as inputs to the system, as taught by Wakeford, and they would have been motivated to do so to allow the AI personal assistant of Osman to automatically determine when the user is preoccupied with and engaging in tasks external to the game application and, in response to such a determination, execute gameplay actions on the user’s behalf, as taught by Wakeford, instead of relying on the less convenient and less efficient method of navigating through a menu and selecting an option to manually activate automatic control of their gameplay by an AI assistant, as taught by Osman, with reasonable expectation of success.
Mousavi teaches a brain-computer interface system for controlling a cursor on a computer screen by employing machine learning to interpret the user’s brain activity. Notably, Mousavi’s system can detect, from the user’s brain activity, whether the system has correctly or incorrectly interpreted the user’s intentions, and adjusts the output of the system accordingly. Thus, Mousavi’s system can be said to employ machine learning to learn a correlation between “a mental state of the user with respect to the input or the output [of the system]” and is capable of changing the output of the system in response to the user’s perception of an erroneous input (“change operation at a time of an erroneous input”). Mousavi also teaches using error-related brain activity to improve the performance of a machine learning classifier in the future, essentially allowing the classifier to “learn from its mistakes” (“A second approach, based on error-driven learning, attempts to limit the occurrence of a future error by updating the classifiers upon the detection of error-related brain activity or to discard the unsupervised adaptation temporarily if an error is detected”). Finally, Mousavi teaches that systems which interpret users’ mental activity based on their brain waves have been known in the art of video games for at least several years at the time of writing (e.g., “For instance, the authors in [28] showed that error signals can be detected from human electrocorticography (ECoG) in a continuous task comprising a video game”).
Elenbaas et al., henceforth Elenbaas, teaches a system for synchronizing sensor data and game event data to influence subsequent game events, wherein the sensor data are physiological data of a user obtained through biosensors. Notably, Elenbaas teaches that temperature may be a useful source of information regarding a user’s mental and/or physiological state while playing a video game (see paragraph 40, “[…] the biofeedback engine 224 may determine a presentation of a subsequent game event, by the game logic 222, to induce the desired level of fear (e.g. […] skin temperature)”).
Osman discloses the use of a head-mounted display as a client device, which includes “Bio-sensors 442 […] to enable detection of physiological data from a user. In one embodiment, the bio-sensors 442 include one or more dry electrodes for detecting bio-electric signals of the user through the user's skin” (paragraph 129, lines 1-5). Dry electrodes, as would be known to one of ordinary skill in the art, are capable of detecting the electrical activity of a human brain when placed on the scalp, and the inclusion of such dry electrodes in a head-mounted display means they would already be in position for recording the user’s brain activity. One of ordinary skill in the art would also be aware that dry electrodes are also capable of detecting the muscle movements of a user by picking up on the electrical signals generated when muscles contract, meaning that Osman’s device would further be inherently capable of detecting the “physical activity” of the user, in conjunction with the cameras, gyroscopes, and accelerometers embedded in or used alongside the head-mounted display (see paragraphs 119-123).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the information processing device of Osman to employ the biosensors already included in the device to record physiological data of the user (such as brain activity, muscle activity, motion data, temperature, etc.) so that the device could provide the mental and physical state of the user as inputs to the machine learning model, as taught by Mousavi, and they would have been motivated to do so to allow Osman’s AI assistant to learn to identify, based on the user’s brain activity, when it has made a mistake while performing autonomous control of the user’s gameplay, as taught by Mousavi, and learn from this feedback to reduce the occurrence of future mistakes, as taught by Mousavi, with reasonable expectation of success.
Claim(s) 5 is/are rejected under 35 U.S.C. 103 as being unpatentable over Osman in view of Mousavi as applied to claim 4 above, and further in view of Warnell et al..
Regarding claim 5, Osman (in view of Mousavi) discloses “the information processing device according to claim 4.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to: replay a play scene of the user, propose the contents of the automatic operation based on the replaced play scene and the feature at the time of the play of the game application; and set the contents of the automatic operation in a form through interaction with the user”.
Having a human being assess the output of a machine learning model and provide feedback to the model based on this assessment is well-known not only in the art of machine learning in general, but also in the art of teaching machine learning models to play video games in particular, as evidenced by Warnell et al. Warnell et al., henceforth Warnell, describes the results of a study exploring using feedback from a human being to drive the behavior of an autonomous agent learning to perform a task in a video game through a framework called TAMER. Warnell concludes that the use of human feedback to train autonomous agents to play a video game is exceedingly lucrative, noting that “After just 15 minutes of real-time interaction with a human, Deep TAMER agents were able to achieve higher scores than agents trained using state-of-the-art deep reinforcement learning techniques and orders of magnitude more training data”.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Osman (in view of Mousavi) to allow users to provide feedback to the system during the process of the AI assistant learning to emulate the user’s gameplay, as taught by Warnell, and they would have been motivated to do so to dramatically improve the performance of the AI assistant, as taught by Warnell, and the results of this modification would have been predictable because the technique of providing human feedback to an autonomous agent learning to play a video game to improve the performance of the autonomous agent had been known in art for several years at the time of the claimed invention, as evidenced by Warnell.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Osman in view of Mousavi as applied to claim 6 above, and further in view of Vange et al. (U.S. Patent No 20220274023).
Regarding claim 7, Osman (in view of Mousavi) discloses “the information processing device according to claim 6.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to enable, based on a trigger, one of interruption of the user operation or cancellation of the user operation, and the trigger is a specific wake word.”
Vange et al., henceforth Vange, teaches a “game engine powered by an artificial intelligence system” that employs an artificial intelligence to automatically perform “low-order” aspects of gameplay in response to “higher order” instructions inferred from inputs issued by the user (see paragraphs 39-41, particularly paragraph 40 lines 1-3, “The artificial intelligence powered game engine 208 provides the player with the ability to perform any task, skill, or action during gameplay with only minimal input”). As Applicant’s and Osman’s inventions are also directed toward the use of artificial intelligence to automate gameplay, Vange is considered to belong to the same field of art. Because Vange teaches the use of voice commands as player inputs to the system (paragraph 36, lines 5-8, “player inputs may comprise active gestures, motions, touch commands, voice commands, or other like actions taken on the user device 210 intended to express an intent during gameplay”), it would have been known in the art at the time of the claimed invention that systems employing artificial intelligence to automate aspects of gameplay could be designed to accept voice commands from a user. It should also be noted that, although Osman is silent on the use of voice commands to initiate or interrupt automated gameplay, the information processing device disclosed by Osman is, at least, inherently capable of receiving vocal commands from a user (see Osman paragraph 125, lines 2-5, “a microphone 422 may be included for capturing audio from the real environment, including sounds from the ambient environment, speech made by the user, etc.” and paragraph 136, lines 13 and 16, “Input devices of Clients 510 may include, for example, […] a voice recognition system”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the information processing device of Osman to allow the user to initiate or interrupt automatic control of the user’s gameplay by the AI assistant by issuing voice commands, as taught by Vange, instead of using an in-game menu, and they would have been motivated to do so to improve the convenience of the user experience. Finally, the results of this modification would have been predictable because not only was Osman’s information processing device already inherently capable of receiving voice commands, the use of voice commands as an input to consumer electronic devices and systems designed for processing video game applications, and in particular systems featuring AI-automated gameplay, was already a known technique in the art at the time of the claimed invention, as evidenced by Vange.
Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over Osman in view of Mousavi as applied to claim 2 above, and further in view of World of Warcraft, as evidenced by Wowpedia and WoW AFK Mode Example Video.
Regarding claim 8, Osman (in view of Mousavi) discloses “the information processing device according to claim 2.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to notify, via a network, a specific user that the user is assisted by the operation assistance processing, and the specific user is different from the user.”
World of Warcraft, a popular massively-multiplayer online game, allows players to enter an “Away/AFK [Away From Keyboard] mode” by entering a command in the in-game chat. When a player enters Away/AFK mode, the word “Away” will appear next to their username in angle brackets, as seen in WoW AFK Mode Example Video and confirmed by Wowpedia. Thus, displaying a visual/textual indicator next to the username of the in-game avatar of an unavailable player in a multiplayer video game to inform other players that the player is not present to control their avatar or respond to in-game messages in cases where their in-game avatar is present in an in-game environment was known in the art at the time of the claimed invention. Furthermore, when a user is in AFK mode, any other user who attempts to send a message to that player receives a notification that the user they are attempting to reach is unavailable, as taught by Wowpedia (“[…] an auto-response will be sent to anyone who sends the flagged user a tell or a ready check”). Because only the specific user who sends the AFK user a message receives this notification, and this notification is not broadcast globally, this reads on the amended limitation of “[…] notify, via a network, a specific user […] and the specific user is different from the user.”
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the information processing device of Osman to include a visual/textual indicator next to the username of a user’s in-game avatar when that user is not presently issuing control inputs to their avatar, as taught by World of Warcraft, including when that user’s avatar is being controlled by an artificial intelligence, as taught by Osman, and they would have been motivated to do so to inform other players wanting to interact with the user that the user is not currently available. Finally, the results of this modification would have been predictable because similar features have been known in the art for several years at the time of the claimed invention, as evidenced by WoW AFK Mode Example Video.
Claim(s) 11-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Osman in view of Mousavi as applied to claim 1 above, and further in view of Borovikov et al. (US Patent No. 20200306643).
Regarding claim 11, Osman (in view of Mousavi) discloses “the information processing device according to claim 1.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to determine, based on the learning result, whether to initiate execution of the operation assistance processing.”
Borovikov et al., henceforth Borovikov, teaches a system for “converting a player-controlled character or virtual entity in a video game to at least temporarily be under emulated control when certain criteria is met.” Borovikov, like Osman’s and Applicant’s inventions, employs a “player imitation model, wherein the player imitation model comprises a trained model configured to receive a current game state of the multiplayer video game as input and to output an in-game action or command based on previously observed player behavior” (paragraph 10). Notably, Borovikov teaches that the imitation model may take control of the player character based on a determination that the observed behavior of the player character does not match the player’s expected behavior for the current gameplay situation (see paragraph 50: “For example, if a player has not moved their virtual character for many seconds in a situation where movement is expected or the player's behavior deviates from expected behavior of the player (e.g., a game state in which the virtual character is being shot at and all of his teammates are running away), the player heartbeat component may indicate a control heartbeat failure to the controller source switch 132. Conversely, there may be another game state, such as a gameplay situation in which it makes sense within the game for the player to not issue any control commands (e.g., the virtual character is hiding from an enemy within the virtual environment), where the player heartbeat component may indicate a consistent control heartbeat despite many seconds of player control inactivity”). Because Borovikov’s “heartbeat protocol” is used to determine when the player imitation model should assume control of the player character (see paragraph 49: “The controller source switch 132 may make the determination of which source should provide the control commands to the application host system based at least in part on a player heartbeat component 136”), the details of the protocol given in paragraph 50 read on the limitation of “the assistance unit is further configured to determine, based on the learning result [the player’s expected behavior], whether to initiate execution of the operation assistance processing [the player imitation model assumes control of the application].”
Finally, Borovikov provides a motivation for switching control to a player imitation model based on a “control heartbeat failure”, that being allowing a player’s gameplay to continue in the event that the player disconnects from a multiplayer game, avoiding player frustration (see paragraph 3: “If one of the players is unable to continue playing at a certain point during a cooperative gameplay session, the remaining players are typically frustrated and at a disadvantage relative to if the player had kept playing (e.g., their team may now only have three actively playing players in a matchup against four competing players on another team)”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of Osman (in view of Mousavi) to include a “control heartbeat” protocol capable of automatically initiating operation assistance of the gaming application, as taught by Borovikov, and they would have been motivated to do so to prevent situations where players may become frustrated or put at a disadvantage in a multiplayer game due to a teammate disconnecting, as taught by Borovikov, with reasonable expectation of success.
Regarding claim 12, Osman (in view of Mousavi) discloses “the information processing device according to claim 1.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to: generate an automatic matching rate indicating a degree of correspondence between the learning result and information related to the situation; and initiate the execution of the operation assistance processing when the automatic operation matching rate satisfies a predetermined criterion.”
Borovikov et al., henceforth Borovikov, teaches a system for “converting a player-controlled character or virtual entity in a video game to at least temporarily be under emulated control when certain criteria is met.” Borovikov, like Osman’s and Applicant’s inventions, employs a “player imitation model, wherein the player imitation model comprises a trained model configured to receive a current game state of the multiplayer video game as input and to output an in-game action or command based on previously observed player behavior” (paragraph 10). Notably, Borovikov teaches that the imitation model may take control of the player character based on a determination that the observed behavior of the player character does not match the player’s expected behavior for the current gameplay situation (see paragraph 50: “For example, if a player has not moved their virtual character for many seconds in a situation where movement is expected or the player's behavior deviates from expected behavior of the player (e.g., a game state in which the virtual character is being shot at and all of his teammates are running away), the player heartbeat component may indicate a control heartbeat failure to the controller source switch 132. Conversely, there may be another game state, such as a gameplay situation in which it makes sense within the game for the player to not issue any control commands (e.g., the virtual character is hiding from an enemy within the virtual environment), where the player heartbeat component may indicate a consistent control heartbeat despite many seconds of player control inactivity”). Because Borovikov’s “heartbeat protocol” is used to determine when the player imitation model should assume control of the player character (see paragraph 49: “The controller source switch 132 may make the determination of which source should provide the control commands to the application host system based at least in part on a player heartbeat component 136”), the details of the protocol given in paragraph 50 read on the limitation of “the assistance unit is further configured to generate an automatic matching rate indicating a degree of correspondence between the learning result and information related to the situation [Borovikov’s system generates a comparison between the player’s observed behavior and their expected behavior given the current gameplay situation]; and initiate the execution of the operation assistance processing when the automatic operation matching rate satisfies a predetermined criterion [the player imitation model assumes control when the player heartbeat component indicates a player control failure].”
Finally, Borovikov provides a motivation for switching control to a player imitation model based on a “control heartbeat failure”, that being allowing a player’s gameplay to continue in the event that the player disconnects from a multiplayer game, avoiding player frustration (see paragraph 3: “If one of the players is unable to continue playing at a certain point during a cooperative gameplay session, the remaining players are typically frustrated and at a disadvantage relative to if the player had kept playing (e.g., their team may now only have three actively playing players in a matchup against four competing players on another team)”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of Osman (in view of Mousavi) to include a “control heartbeat” protocol capable of automatically initiating operation assistance of the gaming application, as taught by Borovikov, and they would have been motivated to do so to prevent situations where players may become frustrated or put at a disadvantage in a multiplayer game due to a teammate disconnecting, as taught by Borovikov, with reasonable expectation of success.
Regarding claim 13, Osman (in view of Mousavi and Borovikov) discloses “the information processing device according to claim 12.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to set contents of the operation assistance processing based on the automatic operation matching rate and the information related to the situation.”
Borovikov et al., henceforth Borovikov, teaches a system for “converting a player-controlled character or virtual entity in a video game to at least temporarily be under emulated control when certain criteria is met.” Borovikov, like Osman’s and Applicant’s inventions, employs a “player imitation model, wherein the player imitation model comprises a trained model configured to receive a current game state of the multiplayer video game as input and to output an in-game action or command based on previously observed player behavior” (paragraph 10). Notably, Borovikov teaches that the player imitation model acts according to how the player would act given the current game context (see paragraph 62: “As discussed, the emulated controls from the controller proxy may be determined based on the current game state and the imitation learning model that may be configured to mimic the player's likely actions in view of the current game state.”), which reads on the limitation of “the assistance unit is further configured to set contents of the operation assistance processing based on the automatic operation matching rate [Borovikov’s comparison between observed vs. expected player behavior] and the information related to the situation [Borovikov’s emulated player behavior is context-dependent].”
Finally, Borovikov provides a motivation for switching control to a player imitation model based on a “control heartbeat failure”, that being allowing a player’s gameplay to continue in the event that the player disconnects from a multiplayer game, avoiding player frustration (see paragraph 3: “If one of the players is unable to continue playing at a certain point during a cooperative gameplay session, the remaining players are typically frustrated and at a disadvantage relative to if the player had kept playing (e.g., their team may now only have three actively playing players in a matchup against four competing players on another team)”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of Osman (in view of Mousavi) to include a “control heartbeat” protocol capable of automatically initiating operation assistance of the gaming application, as taught by Borovikov, and they would have been motivated to do so to prevent situations where players may become frustrated or put at a disadvantage in a multiplayer game due to a teammate disconnecting, as taught by Borovikov, with reasonable expectation of success.
Regarding claim 14, Osman (in view of Mousavi and Borovikov) discloses “the information processing device according to claim 12.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to dynamically enable the operation assistance processing in a case where the automatic operation matching rate satisfies a determined condition.”
Borovikov et al., henceforth Borovikov, teaches a system for “converting a player-controlled character or virtual entity in a video game to at least temporarily be under emulated control when certain criteria is met.” Borovikov, like Osman’s and Applicant’s inventions, employs a “player imitation model, wherein the player imitation model comprises a trained model configured to receive a current game state of the multiplayer video game as input and to output an in-game action or command based on previously observed player behavior” (paragraph 10). Notably, Borovikov teaches that the imitation model may take control of the player character based on a determination that the observed behavior of the player character does not match the player’s expected behavior for the current gameplay situation (see paragraph 50: “For example, if a player has not moved their virtual character for many seconds in a situation where movement is expected or the player's behavior deviates from expected behavior of the player (e.g., a game state in which the virtual character is being shot at and all of his teammates are running away), the player heartbeat component may indicate a control heartbeat failure to the controller source switch 132. Conversely, there may be another game state, such as a gameplay situation in which it makes sense within the game for the player to not issue any control commands (e.g., the virtual character is hiding from an enemy within the virtual environment), where the player heartbeat component may indicate a consistent control heartbeat despite many seconds of player control inactivity”). Because Borovikov’s “heartbeat protocol” is used to determine when the player imitation model should assume control of the player character (see paragraph 49: “The controller source switch 132 may make the determination of which source should provide the control commands to the application host system based at least in part on a player heartbeat component 136”), the details of the protocol given in paragraph 50 read on the limitation of “the assistance unit is further configured to dynamically enable the operation assistance processing in a case where the automatic operation matching rate satisfies a determined condition [the player imitation model assumes control when the player heartbeat component indicates a player control failure].”
Finally, Borovikov provides a motivation for switching control to a player imitation model based on a “control heartbeat failure”, that being allowing a player’s gameplay to continue in the event that the player disconnects from a multiplayer game, avoiding player frustration (see paragraph 3: “If one of the players is unable to continue playing at a certain point during a cooperative gameplay session, the remaining players are typically frustrated and at a disadvantage relative to if the player had kept playing (e.g., their team may now only have three actively playing players in a matchup against four competing players on another team)”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of Osman (in view of Mousavi) to include a “control heartbeat” protocol capable of automatically initiating operation assistance of the gaming application, as taught by Borovikov, and they would have been motivated to do so to prevent situations where players may become frustrated or put at a disadvantage in a multiplayer game due to a teammate disconnecting, as taught by Borovikov, with reasonable expectation of success.
Regarding claim 15, Osman (in view of Mousavi and Borovikov) discloses “the information processing device according to claim 12.”
However, Osman (in view of Mousavi) does not disclose “the assistance unit is further configured to compare the information related to the situation with the learning result to determine the automatic operation matching rate.”
Borovikov et al., henceforth Borovikov, teaches a system for “converting a player-controlled character or virtual entity in a video game to at least temporarily be under emulated control when certain criteria is met.” Borovikov, like Osman’s and Applicant’s inventions, employs a “player imitation model, wherein the player imitation model comprises a trained model configured to receive a current game state of the multiplayer video game as input and to output an in-game action or command based on previously observed player behavior” (paragraph 10). Notably, Borovikov teaches that the imitation model may take control of the player character based on a determination that the observed behavior of the player character does not match the player’s expected behavior for the current gameplay situation (see paragraph 50: “For example, if a player has not moved their virtual character for many seconds in a situation where movement is expected or the player's behavior deviates from expected behavior of the player (e.g., a game state in which the virtual character is being shot at and all of his teammates are running away), the player heartbeat component may indicate a control heartbeat failure to the controller source switch 132. Conversely, there may be another game state, such as a gameplay situation in which it makes sense within the game for the player to not issue any control commands (e.g., the virtual character is hiding from an enemy within the virtual environment), where the player heartbeat component may indicate a consistent control heartbeat despite many seconds of player control inactivity”). Because Borovikov’s “heartbeat protocol” is used to determine when the player imitation model should assume control of the player character (see paragraph 49: “The controller source switch 132 may make the determination of which source should provide the control commands to the application host system based at least in part on a player heartbeat component 136”), the details of the protocol given in paragraph 50 read on the limitation of “the assistance unit is further configured to compare the information related to the situation with the learning result to determine the automatic operation matching rate [Borovikov’s system generates a comparison between the player’s observed behavior and their expected behavior given the current gameplay situation].”
Finally, Borovikov provides a motivation for switching control to a player imitation model based on a “control heartbeat failure”, that being allowing a player’s gameplay to continue in the event that the player disconnects from a multiplayer game, avoiding player frustration (see paragraph 3: “If one of the players is unable to continue playing at a certain point during a cooperative gameplay session, the remaining players are typically frustrated and at a disadvantage relative to if the player had kept playing (e.g., their team may now only have three actively playing players in a matchup against four competing players on another team)”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of Osman (in view of Mousavi) to include a “control heartbeat” protocol capable of automatically initiating operation assistance of the gaming application, as taught by Borovikov, and they would have been motivated to do so to prevent situations where players may become frustrated or put at a disadvantage in a multiplayer game due to a teammate disconnecting, as taught by Borovikov, with reasonable expectation of success.
Response to Arguments
Applicant’s arguments, see pages 11-16, filed 07/29/2026, with respect to the rejection(s) of claim(s) 1-9 under 35 U.S.C. have been fully considered and are persuasive. Therefore, the rejection has been withdrawn.
Applicant's arguments, see pages 1, filed 07/29/2026, with respect to the rejection(s) of claim(s) 1 and 9-10 under 35 U.S.C.2(a)(1) have been fully considered but they are not persuasive. Although the new grounds of rejection do not rely on 35 U.S.C.2(a)(1), Applicant’s arguments regarding the 102(a)(1) rejection are considered relevant to the new grounds of rejection under 35 U.S.C.3, and will be addressed accordingly.
Regarding Osman et al., henceforth Osman, Applicant argues that “Osman is not shown to describe learning a situation by machine learning at a time of a user operation on an application and the situation includes a mental state of the user and a physical state of the user.” Examiner accepts that Osman does not disclose learning the mental and physical state of a user by machine learning. However, the previous version of claim 1 did not in fact reference learning the mental and physical state of the user, so Osman was not required to disclose this feature to read on the limitations of claim 1. Regardless, Examiner respectfully submits that Osman teaches a head-mounted display equipped with various biosensors capable of reading physiological data from the user (see paragraph 129). Thanks to these biosensors, Osman’s device would inherently be capable of learning the mental and physical state of the user as measured by brain activity, heart rate, etc., although Osman is silent on their use for this purpose. However, Mousavi et al. provides a strong motivation to use Osman’s biosensors for this purpose and attests to the use of biosensor technology in the video game field more generally (which is also evidenced by Elenbaas et al.), and has been included as the main secondary reference in the new grounds of rejection under 35 U.S.C.3.
Applicant further argues that “Osman is not shown to describe executing, across at least one application, operation assistance processing to automatically operate the application based on a learning result of the learning situation based on the machine learning.” Examiner respectfully disagrees, and notes that Osman explicitly teaches that the gaming application is operated automatically according to the outputs of a machine learning model trained to emulate a player’s behavior across a variety of game contexts. For example, in paragraph 19, Osman states that “the game play controller can proceed in automatic mode to complete specific game tasks that are difficult for the user.” Osman goes on to specify that “In still another embodiment, the game play controller is configured to play the gaming application after the user pauses the game play, such that an AI character is turned on to perform specific tasks […] In that manner, the game play of the user continues without user attention.” If game play can continue without the user’s attention, then by any reasonable definition this would constitute “automatic” operation of the gaming application.
Finally, Applicant submits that “the portions of Osman cited by the Office Action consistently describe determining how a particular gaming application should be played and controlling gameplay within that application. However, the cited reference Osman is not shown to describe an operation assistance architecture capable of performing operation assistance across multiple applications or independently of a particular application.” Examiner respectfully submits that, as far as the original grounds of rejection were concerned, Applicant’s operation assistance architecture was not positively recited as being “capable of performing operation assistance across multiple applications or independently of a particular application.” Thus, Examiner respectfully submits that the question of whether Osman discloses this limitation was not relevant to the prosecution at the time of the previous office action and, in the case of “[being] capable of performing operation assistance across multiple applications,” continues not to be, as this limitation is not positively recited anywhere in the amended claim set.
Such considerations are moot, however, in light of the fact that Osman’s system is, in fact, both “capable of performing operation assistance across multiple applications” (see paragraphs 34 and 38: “In particular, profiler engine 145 in the game cloud system 210 is configured to build a game play profile for the user that is able to simulate the game play of the user in one or more gaming applications […] through the game cloud system 210, profiler engine 145 is able to collect and analyze game plays of a plurality of users playing a plurality of gaming applications, especially when the game cloud system 210 is executing instances of one or more gaming applications for the plurality of users 115”) and also of “performing operation assistance […] independently of a particular application (see paragraph 37: “ In another embodiment, AI personal assistant 120 may be configured remote to client device 100, and independent of the game cloud system 210”).”
Applicant's arguments, see pages -20, filed 07/29/2026, with respect to the rejection(s) of claim(s) 2-8 under 35 U.S.C.3 have been fully considered but they are not persuasive. These arguments are founded upon Applicant’s assertion that Osman fails to anticipate the recited limitations of the claimed invention, submitting that the various secondary references (Wakeford, Mousavi, Vange, etc.) “[do] not remedy the above-noted deficiencies of Osman,” and additionally noting that “[the dependent claims are] not shown to be taught, suggested, or rendered obvious over the combination of Osman [and the secondary references] based at least on the dependence on amended independent claim 1.” Because the arguments regarding Osman’s deficiencies are considered to be unpersuasive (as discussed above), these arguments, which rest solely on Osman being unsuitable as a primary reference, are similarly considered to be unpersuasive.
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 nonprovisional extension fee (37 CFR 1.17(a)) 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 KENNETH HAROLD JOHANSSON whose telephone number is (571)272-5755. The examiner can normally be reached Monday-Thursday from 8:30 to 6:30.
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, Peter Vasat can be reached at (571)270-7625. 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.
/K.H.J./
Examiner, Art Unit 3715
/WILLIAM H MCCULLOCH JR/Primary Examiner, Art Unit 3715
Prosecution Insights