Chat-based user generated content assistance

Abstract

Embodiments of the present invention include systems and methods for generating chat-based user-generated content. The system may receive a chat-based communication from a user device of a user over a communication network and interpret the chat-based communication using a machine learning model. This may include identifying a meaning of the content in the chat-based communication and predicting that the content in the chat-based communication indicates one or more virtual features. Custom roles for use in a virtual environment may be generated based on the identified meanings and the predicted virtual features. The custom character may include the predicted virtual features. The system may present a custom character to a user via a user interface of a user device.

Description

Technical Field

[0001] This invention generally relates to user-generated content. More specifically, this invention relates to user-generated content assistance using chat-based communication. Background Technology

[0002] Currently available game names and other interactive names allow users to customize objects or characters for use during game sessions and other interactive virtual sessions. For example, customization allows objects or characters to match or otherwise reflect a particular user's personality, likes / dislikes, preferences, mannerisms, appearance, etc. This customization is typically performed based on manual input of desired characteristics selected from a limited set of available traits or other customization factors. For example, various menus and sub-menus (and sub-sub-menus, etc.) may be provided for users to choose from. However, when numerous customization options are available, users may feel overwhelmed or find it difficult to navigate the customization process to find, select, and incorporate the desired customization into digital content. Therefore, customization may require users to invest significant time, thought, and trial and error to customize a character to all their specific needs or desires. Some users (e.g., children) may not be able to manually input desired characteristics and / or may lack the effective communication skills to express how they wish to customize their characters and / or objects.

[0003] While some platforms may offer assistive features and / or content tailored to specific age groups, such features do not address the fundamental problem of simplifying the typically complex and time-consuming customization process for user-specific virtual content (such as characters, objects, and environments). Furthermore, the difficulty of customizing content can lead to an unsatisfactory or poor user experience within the virtual environment, causing users to lose interest in engaging with the content.

[0004] Therefore, there is a need in the art for an improved chat-based user-generated content assistance system and method. Summary of the Invention

[0005] Embodiments of the present invention include a system and method for generating chat-based user-generated content. The system receives chat-based communications from a user's device via a communication network and uses a machine learning model to interpret the chat-based communications. This may include identifying the meaning of content in the chat-based communications and predicting one or more virtual features indicative of the content in the chat-based communications. Custom characters for use in a virtual environment may be generated based on the identified meanings and the predicted virtual features. Custom characters may include the predicted virtual features. The system may present the custom characters to the user via a user interface on the user device. Attached Figure Description

[0006] Figure 1 An exemplary network environment is shown, in which a custom content generation system trained based on machine learning can be implemented.

[0007] Figure 2 A trained custom content generator for generating custom user-generated content is shown according to one implementation scheme.

[0008] Figure 3A This is a flowchart illustrating an exemplary method for chat-based user-generated content assistance according to one implementation scheme.

[0009] Figure 3B This demonstrates how it can be implemented in a custom content generation session. Figure 3A An exemplary user interface for the method.

[0010] Figure 4A This is a flowchart illustrating an exemplary method for adapting custom user-generated content for use in a game, according to one implementation.

[0011] Figure 4B This demonstrates how it can be implemented in a custom content generation session. Figure 4A An exemplary user interface for the method.

[0012] Figure 5 A block diagram of an example electronic entertainment system according to one implementation scheme is shown. Detailed Implementation

[0013] Embodiments of the present invention include a system and method for generating chat-based user-generated content. The system receives chat-based communications from a user's device via a communication network and uses a machine learning model to interpret the chat-based communications. This may include identifying the meaning of content in the chat-based communications and predicting one or more virtual features indicative of the content in the chat-based communications. Custom characters for use in a virtual environment may be generated based on the identified meanings and the predicted virtual features. Custom characters may include the predicted virtual features. The system may present the custom characters to the user via a user interface on the user device.

[0014] Figure 1 An exemplary network environment is illustrated, in which a custom content generation system trained based on machine learning can be implemented. Such a network environment may include a variety of different networking systems and system devices, including a custom content generator 114, a database 116, a user device 122, and a game manager 124. These devices may communicate directly with each other or through one or more intermediate networks (e.g., a local area network, a wide area network, the Internet, a virtual private network, etc.).

[0015] Custom content generator 114 facilitates the generation of user-defined custom user-generated content (e.g., characters, weapons, objects, etc.) associated with user device 122 (e.g., gaming system, laptop computer, desktop computer, smartphone, etc.). As shown, custom content generator 114 may include machine learning model 102, ML model selector 104, ML core process 106, feature extractor 108, historical custom content 110, and custom content manager 112. Such components of custom content generator 114 may be executed by one or more processing devices (e.g., computing devices, mobile devices, servers, databases, etc.) configured to operate together to provide services for custom content generator 114. Components 102 to 112 and processing devices may operate within the same local network (e.g., such as a LAN, WAN, mesh network, etc.) or may be distributed processing devices (e.g., such as a cloud network, distributed processing network, etc.).

[0016] User device 122 can connect to and communicate with custom content generator 114 to trigger a request for a new custom generation session using custom content generator 114. The request may include or otherwise indicate one or more parameters, such as content type (e.g., character, object, etc.), desired game name, and one or more chat-based communications from user device 122. Chat-based communications that can serve as the basis for triggering a custom generation session and a source of custom parameters may include audio clips received via the microphone of user device 122, written communications input to user device 122, any combination thereof, etc. The request may also include other parameters, such as user profile data (associated with user device 122, such as, but not limited to, user characteristics, user demographics, user game history, user preferences, social connections, etc.), previously customized characters and / or content associated with the user, combinations thereof, etc.

[0017] Custom content manager 112 can then instantiate new custom sessions for user device 122. The new custom session can include a specific environment hosted by custom content manager 112 and presented to user device 122 via a graphical user interface. Custom content manager 112, using ML kernel process 106, can provision one or more machine learning models 102 to enable any extended functionality. Machine learning models 102 can be configured to provide natural language processing (e.g., large language models, bidirectional transformers, zero-shot / few-shot learners, deep neural networks, etc.), content generation (e.g., techniques using large language models, deep neural networks, generative adversarial networks, etc.), univariate or multivariate classifiers (e.g., k-nearest neighbors, random forests, log regression, decision trees, support vector machines, gradient descent, etc.), image processing (e.g., using deep neural networks, convolutional neural networks, etc.), sequence data processing (e.g., recurrent neural networks capable of processing datasets organized according to classification sequences, etc.), etc.

[0018] Such machine learning techniques can also be applied to game data and associated user-generated content (UGC), which can be captured during game sessions on different users and user devices 122. This game data can include not only information about the name of the ongoing game but also user profiles, choices, behaviors, etc., associated with the game session. This game data can be monitored and stored in memory as object or activity files, and can be used for supervised and unsupervised learning, thereby training models to identify patterns between game / user data and associated UGC features. In some implementations, the collection of object or activity files can be labeled based on any combination of game metadata and user feedback (including user feedback during content-customized sessions).

[0019] Machine learning model 102 can be trained to process natural language communication (e.g., spoken, textual, etc.) using available user data to generate custom user-generated content based on one or more defined virtual features (e.g., using input from the user, user characteristics, previous custom characters, one or more character parameters of the game name, data associated with one or more additional users, game manager 124, third-party database 116, etc.), generate one or more prompts for the user (e.g., suggesting additional virtual features of custom characters and / or objects, asking follow-up questions after receiving user feedback, etc.), generate custom content (e.g., presenting custom characters, presenting optional features of custom characters, presenting an interface to the user, etc.). Different types of data input can be used to train different machine learning models 102, which may be specific to users, user demographics, associated game or other interactive content names and types, social contacts, etc. Therefore, using the selected data input, machine learning model 102 can be trained to identify information about the requesting user and to identify content-custom parameters that may be explicitly related to the requesting user (e.g., a 10-year-old girl playing a dance game, a 7-year-old boy playing a racing game, and a 31-year-old adult male playing a horror game).

[0020] The ML model selector 104 may be executable to select one or more machine learning models 102 to apply to a request or query received from the user device 122. Such selection by the ML model selector 104 may include: selecting one or more machine learning models 102; generating one or more new machine learning models; or training one or more machine learning models 102 based on user data or associated game data to apply to a request from the user device 122.

[0021] Custom Content Manager 112 can pass chat-based communications (e.g., text and / or audio) extracted through custom sessions, as well as any user data retrieved from or inferred from database 116, to ML Core Process 106 to process the communications using one or more machine learning models selected by ML Model Selector 104. ML Core Process 106 can monitor one or more machine learning models 102 configured to provide communication network services. ML Core Process 106 can train new machine learning models 102, retrain (or boost) existing machine learning models 102, delete machine learning models 102, etc. Because ML Core Process 106 manages the operations of various machine learning models 102, each request to ML Core Process 106 can include identification of a specific machine learning model 102, a requested output, etc., enabling ML Core Process 106 to route the request to the appropriate machine learning model 102 or instantiate and train a new machine learning model 102. Alternatively, ML Core Process 106 can analyze the data to be processed included in the request to select the appropriate machine learning model 102 configured to process that type of data.

[0022] If the ML core process 106 cannot recognize a trained machine learning model 102 configured to handle a request, the ML core process 106 may instantiate and train one or more machine learning models 102 to handle the request. The machine learning model 102 may be trained to process specific inputs and / or generate specific outputs. The ML core process 106 may instantiate and train the machine learning model 102 based on specific data to be processed and / or the specific output requested. For example, user sentiment analysis (e.g., user intent, etc.) can be determined using a natural language processor and / or a classifier, while image processing can be performed using a convolutional neural network.

[0023] The ML core process 106 can select one or more machine learning models 102 based on the features of the data to be processed and / or the expected output. The ML core process 106 can then use a feature extractor 108 to generate a training dataset for the new machine learning model 102 (e.g., in addition to those models configured to perform feature extraction such as some deep learning network). The feature extractor 108 can use historical content 110 to define the training dataset. Historical custom content 110 can store information about features associated with different game options, features associated with previous custom sessions, different combinations thereof, and / or one or more custom characters generated from previous custom sessions. In some cases, previous custom sessions may not involve users of user device 122. Previous custom sessions may include data generated manually and / or procedurally for training machine learning models 102. Historical custom content 110 may not store any information associated with a specific user. Alternatively, historical custom content 110 can store features extracted from custom sessions involving users of user device 122 and / or other users.

[0024] Feature extractor 108 can extract features from historical custom content 110 based on the type of model to be trained and the type of training to be performed (e.g., supervised, unsupervised, etc.). Feature extractor 108 may include search functionality (e.g., procedural search, Boolean search, natural language search, large language model-assisted search, etc.) to enable ML core process 106, administrators, etc., to search for specific datasets in historical custom content 110, thereby improving data selection for training datasets. Feature extractor 108 can aggregate the extracted features into one or more training datasets, which can be used to train the corresponding machine learning model 102 in one or more machine learning models 102. The training datasets may include training datasets for training machine learning models 102, training datasets for testing trained machine learning models, etc. One or more training datasets may be passed to ML core process 106, which can manage the training process.

[0025] Feature extractor 108 can pass one or more training datasets to ML core process 106, and ML core process 106 can initiate the training phase of one or more machine learning models 102. Supervised learning, unsupervised learning, self-supervised learning, etc., can be used to train one or more machine learning models 102. One or more machine learning models 102 can be trained at predetermined time intervals and predetermined number of iterations until one or more target accuracy metrics exceed corresponding thresholds (e.g., accuracy, precision, area under the curve, log loss, F1 score, weighted human inconsistency rate, cross-entropy, mean absolute error, mean squared error, etc.), user input, combinations thereof, etc. Once trained, ML core process 106 can use additional training datasets to validate and / or test the trained machine learning models 102. Machine learning models 102 can also be trained at runtime using reinforcement learning.

[0026] Once the machine learning model 102 is trained, the ML core process 106 can manage the operation of one or more machine learning models 102 (stored together with other machine learning models 102) during runtime. The ML core process 106 can instruct the feature extractor 108 to define feature vectors from received data (e.g., audio clips from user device 122, chat-based communications from user device 122, responses from user device 122 to prompts made by custom content generator 114, etc.). In some cases, the ML core process 106 can facilitate the generation of feature vectors whenever the communication channel changes (e.g., transmitting audio clips from user device 122 via the communication channel, receiving new chat-based communications, receiving responses to prompts from user device 122, receiving data related to a specific game name from game manager 124, etc.). The ML core process 106 can continuously execute one or more machine learning models 102 to generate corresponding outputs. The ML core process 106 can evaluate the output to determine whether to manipulate the user interface (e.g., and / or virtual reality interface) of a custom session based on the output (e.g., publishing generated custom characters and / or objects, presenting generated prompts to the user to obtain feedback, modifying custom characters and / or objects according to a specified game name, presenting additional custom character and / or object features to the user, etc.).

[0027] For example, the ML core process 106 can detect new audio segments in a chat-based communication within a custom session. The ML core process 106 can execute a machine learning model 102 (e.g., a recurrent neural network) to process the audio segments to determine words (if any) and sentiment (e.g., the predicted meaning of individual words or words as a whole) within the audio segments. The ML core process 106 can execute another machine learning model 102 (e.g., a classifier, a large language model, and / or a transformer, a generative adversarial network, etc.) to generate content corresponding to the words and / or sentiment, which can be provided to the user device 122. For example, the words could include "Make me an orange monster" with the sentiment of "characteristics of a custom character." The ML core process 106 can also execute another machine learning model 102 to identify user (e.g., a child versus an adult) characteristics based on speech analysis, image analysis (e.g., analysis of user images captured by the user device 122's camera), user profile data, etc., thereby determining whether the "orange monster" should be generated with a childlike (as opposed to a terrifying) appearance, voice, abilities, in-game actions, etc. Another machine learning model 102 can process words and emotions for the user interface (e.g., Figure 5 The user interface 518 and / or virtual reality interface 538 generate content such as one or more custom characters regarded as “orange monsters”, one or more suggestions for additional characteristics of the custom characters (e.g., weapons, accessories, wings, color attributes, body attributes, etc.), one or more prompts for additional characteristics designed to collect user feedback (e.g., “Your orange monster can’t walk, what can help him walk?”), etc.

[0028] The ML core process 106 can instruct the feature extractor 108 to define additional feature vectors to use the machine learning model 102 to process other data in parallel with the aforementioned machine learning model 102, thereby providing additional resources for the custom content generator 114. The ML core process 106 can execute any number of machine learning models in parallel to provide functionality for custom sessions.

[0029] Figure 2 A trained custom content generator 114 for generating custom user-generated content according to one embodiment is shown. The custom content generator 114 can receive input from a game manager 124, user input 202, historical custom content 110, user characteristics 204, a third-party database 116, and any combination thereof, to generate custom content to be presented to a user device 122.

[0030] Custom content generator 114 may utilize one or more different data sources for training, generating, initializing, instantiating, and / or utilizing one or more machine learning models trained to generate custom user-generated content for use in one or more game sessions associated with one or more game names. In some cases, custom content generator 114 may receive input from game session manager 124. For example, custom content generator 114 may receive game names from user device 122. Custom content generator 114 may query game session manager 124 for one or more character parameters required for a specific game name. Character parameters may be custom character features required for valid game sessions for a specific game name. For example, a game name may require a character to fly, so character parameters may indicate that a custom character needs wings, a jetpack, magical abilities, a hoverboard, etc. Custom content generator 114 may generate custom content based on the character parameters associated with the game name.

[0031] In some examples, character parameters may be incompatible with other features of the custom content (e.g., the game might require legs, but the custom character might be a snake). The custom content generator 114 may generate one or more notifications and / or prompts for the user to inform them that the custom content is incompatible with a specific game name. These notifications and / or prompts may be warning messages (e.g., "Your custom character cannot play the game name"), prompts for adding / removing features (e.g., "The game name requires a lot of running; would you like to add legs to your custom character?"), heuristic questions designed to guide the user (e.g., "The game name seems to require your character to fly. What can help your character fly?"), notifications that features have been added (e.g., "Eyes have been added to your character to play the game name."), or any combination thereof. Upon receiving a notification and / or prompt, the user device 122 may output one or more messages to the custom content generator 114, indicating approval or disapproval (if necessary) of adjustments made to the custom character. The custom content generator 114 can also adapt custom characters in different ways for different game names, including changing the color palette, animation or graphic art style, clothing or accessories, equipment or weapons, etc., to coordinate with the art or design elements in the specific interactive content name.

[0032] Custom content generator 114 can receive input from user input 202. User input 202 may include chat-based communication from a user. Chat-based communication may include audio clips (e.g., commands received via a microphone associated with user device 122), text communication (e.g., messages sent to custom content generator 114 via user device 122), gestures (e.g., body movements or gestures received via visual input (such as a camera or webcam) associated with user device 122), any combination thereof, etc. Custom content generator 114 may employ one or more machine learning models (e.g., Figure 1 The machine learning model discussed in the text determines the meaning of chat-based communication and / or predicts one or more virtual features of a custom object. For example, a user might enter a chat-based communication stating "I want a pink dragon," and the first machine learning model can interpret the meaning of the chat-based communication and one or more predicted virtual features. In this example, the meaning might be "desired features of a custom character," and some predicted virtual features might be "pink" and "dragon."

[0033] Custom content generator 114 can receive input from historical custom content 110. Historical custom content 110 may include previously generated custom content associated with the user and / or other users. For example, if a user has previously shown a tendency to request a specific virtual characteristic, custom content generator 114 can automatically add that specific virtual characteristic to the custom object. In some examples, if a specific virtual characteristic might be popular among users with similar demographic characteristics to the user, custom content generator 114 can automatically add said specific virtual characteristic to the custom object. Other data from historical custom content 110 used by custom content generator 114 may include certain combinations of virtual characteristics, colors, design styles (e.g., cartoon, realistic, etc.), clothing / accessories / hairstyles, two-dimensional or three-dimensional elements, or any combination thereof. In some examples, the user may instruct editing of previously generated custom characters and / or content. Custom content generator 114 can query previously generated custom characters and / or content in historical custom content 110 based on input from the user and perform adjustments based on communications from the user.

[0034] Custom content generator 114 can also receive input from user features 204. User features 204 may include one or more user features associated with the user, such as age, demographic characteristics, appearance, preferences, language, geographic region, and any combination thereof. Custom content generator 114 can generate one or more virtual features based on one or more user features received from user features 204. For example, if the user has blonde hair, custom content generator 114 can generate a custom character with blonde hair. In another example, if the user is from a specific geographic region, custom content generator 114 can generate a custom character with similar speech, dialect, and / or language. User features 204 can query one or more user features from a user profile associated with the user and / or user device 122. The user profile may be stored on user device 122 and may contain user input data, generated user-related metadata, and / or predictions about the user. User profiles may also include data collected during sessions with one or more games associated with user device 122 and the user (e.g., when the user is playing a specific game name, during a custom session, when the user participates in a chat room hosted by a central server accessible to user device 122, etc.). User characteristics may also include direct social circles (e.g., online friends, teammates), and any other players or content the user may follow or has subscribed to, as indicated by user activity data. Thus, users may also express queries about their favorite players or characters (e.g., “I want a character like BigNameStreamer27’s character in game name 6, but with blue hair instead of green”).

[0035] The custom content generator 114 can receive input from one or more third-party databases 116. The third-party databases 116 may include data provided by one or more sponsors, data from a central server, data from one or more programs running simultaneously on the user device 122 (e.g., parental controls), data from additional demographic data sources (e.g., social media), or any combination thereof. The provided data may include review parameters determined by parental controls, administrative controls, settings associated with the user device 122, game names, or any combination thereof.

[0036] Custom content generator 114 can receive data from database 116 and generate one or more virtual features based on the received data. For example, if a third party is a sponsor of the game name and / or the customization process facilitated by custom content generator 114, custom content generator 114 can place a logo on the custom content (e.g., place a logo on a custom character's T-shirt, hat, or bag). As another example, parental controls may restrict user device 122 from adjusting custom characters that may be considered obscene.

[0037] In some examples, the custom content generator 114 can connect to one or more social media websites associated with the user. The custom content generator 114 can generate one or more virtual features based on content posted on the social media websites associated with the user. For example, if the user recently posted photos of a rock concert, the custom content generator 114 might generate one or more virtual features that mimic the rock star (e.g., spiky hair, guitar, tattoos, etc.).

[0038] In some implementations, the custom content generator 114 can generate unique game names and / or unique game sessions, which are customized based on user and / or user-generated custom content. The custom content generator 114 can receive input from the game manager 124 and / or the user device 122, which may include the user's most frequently played game names, the user's likes / dislikes / preferences, the user's age range, the user's skills, etc. The custom content generator 114 can employ one or more machine learning models trained to generate custom content (e.g., using large language models, deep neural networks, generative adversarial networks, etc.). For example, the custom content generator 114 can employ a second machine learning model to generate one or more custom interactive storylines including custom characters. The custom interactive storylines can be compatible with one or more game names, or can be unique game names accessible to the user device 122.

[0039] Custom content generator 114 can output custom content during and / or after the duration of a custom session with the user. Custom content may include one or more virtual features generated from one or more sources described herein (e.g., game mode manager 124, user input 202, historical custom content 110, user characteristics 204, database 116, any combination thereof, etc.). Custom content may also include unique game sessions, unique game names, and / or custom interactive storylines generated by custom content generator 114. Custom content can be output to the communication interface of user device 122 (e.g., ...). Figure 5User interface 518 and / or virtual reality interface 538).

[0040] Figure 3A This is a flowchart illustrating an exemplary method 300 for chat-based user-generated content assistance according to one embodiment. Although the example flowchart depicts a specific sequence of operations, this sequence can be changed without departing from the scope of this disclosure. For example, some of the depicted operations may be performed in parallel or in a different order that does not substantially affect the functionality of the flowchart. In other examples, different components of the example apparatus or system implementing the flowchart may perform functions substantially simultaneously or in a specific order.

[0041] Method 300 may include receiving chat-based communication from a user's user device 122 via a communication network at block 302. Chat-based communication may be audio clips received via a microphone associated with the user, written communication input to the user device 122, any combination thereof, etc. The request may also include other parameters, such as user profile data (associated with the user, such as, but not limited to, user characteristics, user demographics, user game history, etc.), previously customized roles and / or content associated with the user, combinations thereof, etc.

[0042] Custom Content Manager 112 (for example, Figure 1 The custom content manager 112 described herein can instantiate a new custom session for user device 122. The custom session may include a specific environment hosted on user device 122 and presented via a communication interface. The appearance and characteristics of the custom session and / or the specific environment may depend on one or more user characteristics derived from the user profile. For example, if the user is young, the custom session may include simplified images and commands.

[0043] According to some examples, method 300 may also include using a machine learning model at box 304 to interpret chat-based communication, wherein interpreting chat-based communication includes identifying the meaning of content in chat-based communication and predicting one or more dummy features indicating the content in chat-based communication. Custom content manager 112 may provision one or more machine learning models to interpret chat-based communication, such as determining the user's sentiment (e.g., the predicted meaning of individual words or words as a whole), the user's intent (e.g., the purpose of individual words or words as a whole), and one or more predicted dummy features of the custom content. One or more machine learning models may be trained to interpret natural language (e.g., recurrent neural networks).

[0044] Method 300 may further include generating a custom persona for a virtual environment at box 306 based on the identified meaning and predicted virtual features, wherein the custom persona includes the predicted virtual features. Custom content manager 112 may be configured with one or more machine learning models to generate custom content for the user using an interpretation of chat-based communication. For example, custom content manager 112 may employ machine learning models (e.g., classifiers, large language models and / or transformers, generative adversarial networks, etc.) trained to generate content corresponding to the user's words and / or sentiments.

[0045] Custom content generated by a machine learning model can include predicted virtual features, as well as additional features generated from inputs from one or more data sources. For example, the machine learning model may receive inputs about historical custom content 110, user characteristics, one or more character parameters associated with a game name, and / or one or more additional third-party databases. One or more data sources can provide one or more additional virtual features applicable to the custom content. For example, historical custom content 110 and / or user characteristics might indicate that other users with similar demographic characteristics prefer blue, so the machine learning model might generate custom content that is blue. In another example, a specific game name might indicate that a playable character must be able to walk, so the machine learning model might generate a custom character with legs (e.g., instead of a snake, worm, etc.).

[0046] According to some examples, the method includes displaying a custom character in the user interface of user device 122 at box 308. Custom content manager 112 can present custom content to the user within the specific context of a custom session associated with user device 122. This specific context can be presented on a user interface such as a television, screen, virtual reality interface, laptop computer, desktop computer, or any combination thereof.

[0047] In some examples, the rendering of custom characters and / or objects may be accompanied by one or more prompts from the Custom Content Manager. Prompts may include warning messages (e.g., "Your custom character cannot play Game Name," "Your custom weapon cannot be used in Game Name"), prompts for adding / removing features (e.g., "Game Name requires a lot of running, do you want to add legs to your custom character?"), heuristic questions designed to guide the user (e.g., "Game Name seems to require your character to fly. What can help your character fly?"), notifications to the user of added features (e.g., "Eyes added to your character to play Game Name," "Shield added to match your custom sword"), or any combination thereof. Upon receiving a prompt, user device 122 may output one or more messages to Custom Content Manager 112 indicating approval / disapproval of adjustments made to the custom content, additional virtual features to be added to the custom content, adjustments to the custom content, any combination thereof, etc.

[0048] When the custom content manager receives one or more messages, one or more machine learning models can interpret those messages (e.g., as specified in box 304) and regenerate the custom content (e.g., as specified in box 306). The custom content can then be re-presented on user device 122 after one or more adjustments.

[0049] Figure 3B This demonstrates how it can be implemented in a custom content generation session. Figure 3A An exemplary user interface for the method. The user interface can be displayed via user device 122 during the duration of a custom session. A user can input chat-based communications via one or more user devices 122 (e.g., keyboard, microphone, mobile device, controller, etc.), wherein the chat-based communications can be displayed in chat interface 314 when received by a communication network. Chat interface 314 can display one or more responses and / or one or more prompts generated by a machine learning model configured to interpret natural language from the user. Figure 3B As shown, the user is in a custom session used to customize a custom persona 312 within a custom interface 310. After receiving chat-based communication from the user, the custom content generator 114 can update the custom persona 312 based on the chat-based communication. For example, as... Figure 3B As shown, a user can request a custom character 312 similar to the user. Therefore, the custom content generator 114 may provide a custom character 312 with short hair, brown hair, brown eyes, and fair skin.

[0050] Figure 4AThis is a flowchart illustrating an exemplary method for adapting custom user-generated content for game performance according to one embodiment. Although the example flowchart depicts a specific sequence of operations, this sequence can be changed without departing from the scope of this disclosure. For example, some of the depicted operations may be performed in parallel or in a different order that does not substantially affect the functionality of the flowchart. In other examples, different components of the example apparatus or system implementing the flowchart may perform functions substantially simultaneously or in a specific order.

[0051] Method 400 may also include displaying a custom character in the user interface of user device 122 at box 402. Custom content manager 112 may present custom content (e.g., custom characters and / or objects) to the user within a specific context of a custom session associated with user device 122. This specific context may be presented on a user interface such as a television, screen, virtual reality interface, laptop computer, desktop computer, or any combination thereof.

[0052] As shown in the figure, method 400 may further include receiving an option selection from user device 122 at box 404, the option corresponding to using a custom character to participate in a game session with a game name. The user can specify a specific game name to utilize the custom character. The option selection can be received via chat-based communication, which may be audio clip input to user device 122 capable of receiving audio clips, and / or text-based communication input to user device 122 capable of receiving text / words. Custom content manager 112 can input the chat-based communication into one or more machine learning models (e.g., Figure 1 The machine learning models discussed herein are trained to interpret natural language. One or more machine learning models can determine the game name associated with the chat-based communication by searching a game history database associated with user device 122, which contains game history data. Game history may include games downloaded on user device 122, most frequently played games on user device 122, games historically played on user device 122, games browsed in online marketplaces associated with user device 122, and any combination thereof. Based on the database, one or more machine learning models can identify a specific game name indicated by user device 122 via chat-based communication.

[0053] According to some examples, method 400 may also include querying the game manager at box 406 for one or more character parameters associated with the game name. The game manager (e.g., Figure 1 and Figure 2The game manager 124 described herein can receive and / or store one or more character parameters associated with a specific game name. For example, after selection by user device 122, the game manager can query a database associated with a specific game name for one or more character parameters associated with that specific game name. In some examples, the game manager may automatically store one or more character parameters based on game names included in the game history database.

[0054] Character parameters may be characteristics of custom characters and / or objects that are necessary to utilize the character in a specific game title. For example, if a specific game title primarily takes place underwater, character parameters might include underwater breathing methods (e.g., gills, scuba gear, magical abilities, etc.) and / or swimming methods (e.g., fins, flippers / fins, tail, etc.). As another example, a specific game title may require running and jumping, so character parameters might include running methods (e.g., two legs, four legs, wheels, etc.) and / or jumping methods (e.g., two legs, four legs, jetpack / propulsion system, etc.). Character parameters may be included in the data packet received when downloading a specific game title. In some examples, the game manager may employ one or more machine learning models (e.g., Figure 1 The machine learning models discussed herein are trained to identify commonalities among characters in a specific game title. These machine learning models can identify character parameters and output them to the game manager.

[0055] According to some examples, method 400 may also include verifying at box 408 whether the custom character includes character parameters associated with the game name, wherein the game session is initiated based on said verification. Custom content manager 112 may verify whether the custom character includes character parameters identified by the game session manager. In some examples, custom content manager 112 may employ one or more machine learning models to identify features of the custom character. For example, one or more machine learning models may determine that the custom character has arms, wings, a head, eyes, a nose, and weapons. Custom content manager 112 may cross-reference the character parameters with the determined features of the custom character (e.g., arms, wings, head, eyes, nose, and weapons) and may determine that the custom character contains all the character parameters necessary to complete a game session for a specific game name. User device 122 may receive a verification notification, and the game session may be initiated immediately.

[0056] Method 400 may also include initially determining at box 410 that the custom character does not include character parameters associated with the game name. In some examples, the custom content manager 112 may cross-reference one or more character parameters with identified features of the custom character (e.g., arms, wings, head, eyes, nose, and weapons), and may determine that the custom character does not have all the character parameters necessary to complete a session for a specific game name. For example, a custom character may have arms, wings, head, eyes, nose, and weapons, and one or more character parameters may include "legs" and "mouth".

[0057] In some examples, one or more character parameters can be assigned values ​​corresponding to a necessity level. These values ​​can be from 0 to 1, where 0 corresponds to "not necessary to complete the game session" and 1 corresponds to "extremely necessary to complete the game session". In some examples, a threshold can be set by a parameter for a specific game name, user device 122, administrator, etc., indicating the minimum value necessary to adjust the custom character. For example, the threshold can be set to 0.4, which indicates that all character parameters with values ​​less than 0.4 are not necessary to complete the game session. All character parameters with values ​​greater than or equal to 0.4 may be necessary to complete the game session. As an additional example, a custom character may have arms, wings, a head, eyes, a nose, and weapons, and one or more character parameters may include a "leg" with a value of 0.9 and a "mouth" with a value of 0.2. Assuming a threshold of 0.4, user device 122 can adjust the custom character to include the legs but not the mouth, and can continue to complete the game session.

[0058] According to some examples, the method includes using a machine learning model at box 412 to generate one or more character modifications, wherein the one or more character modifications conform to character parameters. In some examples, the custom character manager may employ one or more machine learning models to generate additional virtual features for the custom character based on one or more character parameters. In some examples, the custom character manager may notify the user device 122 of any modifications via one or more prompts. Prompts may include warning messages (e.g., "Your custom character cannot play Game Name"), prompts for adding / removing features (e.g., "Game Name requires a lot of running, would you like to add legs to your custom character?"), heuristic questions designed to guide the user (e.g., "Game Name seems to require your character to fly. What can help your character fly?"), notifications to the user of added features (e.g., "Eyes have been added to your character to play Game Name."), or any combination thereof, etc.

[0059] According to some examples, the method includes applying role modifications to a custom role at box 414, where the custom role is verified based on the role modifications. Upon receiving a prompt, user device 122 can output one or more messages to custom content manager 112, indicating approval / disapproval of adjustments made to the custom content, additional virtual features to be added to the custom content, adjustments to the custom content, any combination thereof, etc. Custom content manager 112 can update the custom role accordingly.

[0060] The Custom Content Manager 112 can verify whether a custom character includes character parameters identified by the Game Play Manager. In some examples, the Custom Content Manager 112 can employ one or more machine learning models to identify features of the custom character. For example, one or more machine learning models can determine that the custom character has arms, wings, a head, eyes, a nose, and weapons. The Custom Content Manager 112 can cross-reference one or more character parameters with the identified features of the custom character (e.g., arms, wings, head, eyes, nose, and weapons) and can determine that the custom character contains all the character parameters necessary to complete a Game Play session for a specific game name. The user device 122 may receive a verification notification, and the Game Play session may start immediately. In some examples, the custom character may not include one or more character parameters. The Custom Character Manager can re-emulate one or more machine learning models to determine one or more character parameters that must be added to the custom character for use in the Game Play session. The Custom Character Manager can replicate the methods and systems disclosed herein.

[0061] Figure 4B This demonstrates how it can be implemented in a custom content generation session. Figure 4A An exemplary user interface for the method. During the duration of a custom session, the exemplary interface can be displayed via user device 122. A user can input chat-based communications via one or more user devices 122 (e.g., keyboard, microphone, mobile device, controller, etc.), wherein the chat-based communications can be displayed in chat interface 420 when received by a communication network. Chat interface 420 can display one or more responses and / or one or more prompts generated by a machine learning model configured to interpret the user's natural language. Figure 4B As shown, the user is in a custom session used to customize a custom role 418 within a custom interface 416. After receiving chat-based communication from the user, the custom content generator 114 can update the custom role 418 based on the chat-based communication. For example, as... Figure 4BAs shown, the custom content generator 114 can notify the user that the custom character 418 may require one or more character parameters to be used during a session in a game of a specific game name. The custom content generator 114 can also prompt the user via chat-based communication within the chat interface 420 that the custom character 418 may require one or more modifications for a specific game name.

[0062] Figure 5 A block diagram of an exemplary electronic entertainment system 500 according to an embodiment of the present invention is shown. Figure 5 The illustrated electronic entertainment system 500 includes a main memory 502, a central processing unit (CPU) 504, a graphics processor 506, an input / output (I / O) processor 508, a controller input interface 510, a hard disk drive or other storage component 512 (which may be removable), a communication network interface 514, a virtual reality interface 516, a sound engine 518, and an optical disc / media control device 520. Each of the above is connected via one or more system buses 522.

[0063] like Figure 5 The illustrated electronic entertainment system 500 may be a video game console. Alternatively, the electronic entertainment system 500 may be implemented as a general-purpose computer, set-top box, handheld gaming device, tablet computing device, mobile computing device, or mobile phone. The electronic entertainment system may include some or all of the disclosed components, depending on specific form factors, purposes, or designs.

[0064] Main memory 502 stores instructions and data for execution by CPU 504. When the electronic entertainment system 500 is in operation, main memory 502 can store executable code. Figure 5 The main memory 502 can communicate with the CPU 504 via a dedicated bus. In addition to programs transferred from the hard disk drive / storage component 512, DVD or other optical disc (not shown) via the I / O processor 508 using the optical disc / media controller 520 or downloaded via the communication network interface 514, the main memory 502 can also provide pre-stored programs.

[0065] Figure 5 The graphics processor 506 (or graphics card) executes graphics instructions received from the CPU 504 to generate images for display on a display device (not shown). Figure 5The graphics processing unit (GPU) 506 can convert objects from three-dimensional coordinates to two-dimensional coordinates and vice versa. GPU 506 can use ray tracing to help render light and shadow in game scenes by simulating and tracing individual rays generated from sources. GPU 506 can take advantage of fast startup and loading times, 4K-8K resolution, and up to 120 FPS and 120 Hz refresh rates. GPU 506 can render or process images differently for specific display devices.

[0066] Figure 5 The I / O processor 508 can also allow content to be exchanged via wireless or other communication networks, such as IEEE 802.x, 5G, 4G, LTE and 3G mobile networks including Wi-Fi and Ethernet, as well as Bluetooth and short-range personal area networks. Figure 5 The I / O processor 508 primarily controls the data exchange between various devices of the electronic entertainment system 500, including a CPU 504, a graphics processor 506, a controller interface 510, a hard disk drive / storage component 512, a communication network interface 514, a virtual reality interface 516, a sound engine 518, and an optical disc / media control 520.

[0067] Figure 5 The user of the electronic entertainment system 500 provides instructions to the CPU 504 via a controller device communicatively coupled to the controller interface 510. Various controllers can be used to receive instructions, including handheld controllers and sensor-based controllers (e.g., for capturing and interpreting eye-tracking-based commands, voice-based commands, and gesture commands). The controller can receive instructions or input from the user, and then provide those instructions or input to the controller interface 510, and then to the CPU 504 for interpretation and execution. The CPU 504 can also use these instructions to control other components of the electronic entertainment system 500. For example, the user can instruct the CPU 504 to store certain game information on the hard disk drive / storage component 512 or other non-transitory computer-readable storage medium. The user can also instruct a character in the game to perform a specific action, which is rendered by the graphics processor 506, including audio interpreted by the sound engine 518.

[0068] The hard disk drive / storage component 512 may include a removable or non-removable non-volatile storage medium. The Saud medium may be portable and includes digital video optical discs, Blu-ray discs, or USB-coupled storage devices for inputting and outputting data and code to and from the main memory 502. Software used to implement embodiments of the invention may be stored on such media and input to the main memory via the hard disk drive / storage component 512. The software stored on the hard disk drive 512 may also be managed via the optical disc / media control 520 and / or the communication network interface 514.

[0069] The communication network interface 514 can allow communication via various communication networks, including local private networks and / or larger wide area networks (such as the Internet). The Internet is a widespread network interconnecting computers and servers, allowing Internet Protocol (IP) data to be transmitted and exchanged between users connected through a network service provider. Examples of network service providers include: public switched telephone networks, wired or fiber optic service providers, digital subscriber lines (DSL) or broadband, and satellite services. The communication network interface allows the exchange of communication and content between various remote devices, including other electronic entertainment systems associated with other users and cloud-based databases, services, and servers, as well as content hosting systems that may provide or facilitate gameplay and related content.

[0070] The virtual reality interface 516 allows for the processing and rendering of virtual reality, augmented reality, and mixed reality data. This includes display devices, enabling partially or fully immersive virtual environments. The virtual reality interface 516 can allow for the exchange and presentation of immersive vision and foveated rendering in coordination with sound and haptic feedback processed by the sound engine 518.

[0071] The sound engine 518 executes instructions to generate sound signals, which are output to audio devices such as television speakers, controller speakers, stand-alone speakers, headphones, or other headphone speakers. Each of the different sound output devices can produce different sets of sounds. This may include spatial or three-dimensional audio effects.

[0072] The optical disc / media control 520 can be implemented using a magnetic disk drive or an optical disc drive for storing, managing, and controlling data and instructions for use by the CPU 504. The optical disc / media control 520 may include system software (operating system) for implementing embodiments of the present invention. This system facilitates loading the software into main memory 502.

[0073] The foregoing detailed description of the present technology has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the technology to the precise forms disclosed. In view of the above teachings, many modifications and variations are possible. The described embodiments have been chosen to best explain the principles of the present technology, its practical application, and to enable others skilled in the art to utilize the technology in various embodiments and make various modifications suitable for the intended particular purpose. The scope of the present technology is intended to be defined by the claims. Claims (as amended under Article 19 of the Treaty) 1. A computer-implemented method for creating user-generated content, the method comprising: Receive chat-based communications from the user's device via a communication network; Using a machine learning model to interpret the chat-based communication, wherein interpreting the chat-based communication includes identifying the meaning of the content in the chat-based communication based on the user's user demographic data and predicting that the content in the chat-based communication indicates one or more virtual features; A custom role for use in a virtual environment is generated based on the identified meaning, the user's user demographic data, and the predicted virtual characteristics, wherein the custom role includes the predicted virtual characteristics; and The custom character is displayed in the user interface of the user device. 2. The computer-implemented method of claim 1, wherein the user interface includes one or more options for using the custom character, and further includes: The user device receives an option selection, the option corresponding to using the custom character to participate in a game session with the game name; Query the game manager for one or more character parameters associated with the game name; and Verify whether the custom character includes the character parameters associated with the game name, wherein the game session is initiated based on the verification. 3. The computer-implemented method as described in claim 2, further comprising: It has been preliminarily determined that the custom character does not include the character parameters associated with the game name; The machine learning model is used to generate one or more character modifications, wherein the one or more character modifications conform to the character parameters; and The character modification is applied to the custom character, wherein the custom character is verified based on the character modification. 4. The computer-implemented method of claim 1, further comprising using the machine learning model to generate one or more custom interactive storylines including the custom character, wherein the custom interactive storylines are presented in the virtual environment. 5. The computer-implemented method of claim 4, wherein the one or more custom interactive storylines are generated based on feedback from the user device. 6. The computer-implemented method of claim 1 further includes verifying whether the one or more content features conform to a set of audit parameters associated with the user's user demographic data. 7. The computer-implemented method of claim 1, wherein the custom role is one of a set of different custom roles generated for including the virtual features and presented in the user interface, and wherein the user interface also presents a prompt to select one of the set of different custom roles. 8. The computer-implemented method of claim 1, further comprising generating design prompts to present to the user device, wherein the chat-based communication is received in response to the design prompts. 9. The computer-implemented method of claim 1, wherein the machine learning model is trained using historical game data about stored characters, including custom characters. 10. The computer-implemented method of claim 9, wherein the machine learning model is further trained using associated data from at least one of one or more additional users, user characteristics, game names, or game features. 11. The computer-implemented method of claim 1, wherein the custom role includes predicted virtual features, and wherein the custom role includes predicted virtual features associated with one or more stored custom roles. 12. The computer-implemented method as described in claim 1, further comprising: Identify one or more user characteristics associated with the user; and Using the machine learning model, additional virtual features are generated based on the identified user characteristics, wherein the generation of the custom role for use in the virtual environment is also based on the identified user characteristics. 13. The computer-implemented method of claim 1, further comprising generating one or more prompts, the one or more prompts presenting one or more additional virtual feature options selectable to be applied to the custom character. 14. A computing device, comprising: A communication interface that communicates with a user's device via a communication network, wherein the communication interface receives chat-based communications from the user device; and Processor, the processor executing instructions stored in memory, wherein the processor executing the instructions to: Machine learning models are used to interpret the chat-based communications, wherein interpreting the chat-based communications includes identifying the meaning of the content in the chat-based communications based on the user's user demographic data and predicting that the content in the chat-based communications indicates one or more virtual features; and A custom role for use in a virtual environment is generated based on the identified meaning, the user's user group characteristic data, and the predicted virtual characteristics, wherein the custom role includes the predicted virtual characteristics, and the custom role is displayed in the user interface of the user device. 15. The computing device of claim 14, wherein the user interface includes one or more options for using the custom character, and wherein the processor executes further instructions to: The user device receives an option selection, the option corresponding to using the custom character to participate in a game session with the game name; Query the game manager for one or more character parameters associated with the game name; and Verify whether the custom character includes the character parameters associated with the game name, wherein the game session is initiated based on the verification. 16. The computing device of claim 15, wherein the processor executes further instructions to: It has been preliminarily determined that the custom character does not include the character parameters associated with the game name; The machine learning model is used to generate one or more character modifications, wherein the one or more character modifications conform to the character parameters; and The character modification is applied to the custom character, wherein the custom character is verified based on the character modification. 17. The computing device of claim 14, wherein the processor executes further instructions to generate one or more prompts for the user, wherein the prompts include one or more additional virtual features applicable to the custom role that the user may approve or reject. 18. A non-transitory computer-readable storage medium, the computer-readable storage medium comprising instructions that, when executed by a computer, cause the computer to: Receive chat-based communications from the user's device via a communication network; Using a machine learning model to interpret the chat-based communication, wherein interpreting the chat-based communication includes identifying the meaning of the content in the chat-based communication based on the user's user demographic data and predicting that the content in the chat-based communication indicates one or more virtual features; A custom role for use in a virtual environment is generated based on the identified meaning, the user's user demographic data, and the predicted virtual characteristics, wherein the custom role includes the predicted virtual characteristics; and The custom character is displayed in the user interface of the user device.

Claims

1. A computer-implemented method for creating user-generated content, the method comprising: Receive chat-based communications from the user's device via a communication network; Using a machine learning model to interpret the chat-based communication, wherein interpreting the chat-based communication includes recognizing the meaning of the content in the chat-based communication and predicting that the content in the chat-based communication indicates one or more virtual features; A custom role for use in a virtual environment is generated based on the identified meaning and the predicted virtual features, wherein the custom role includes the predicted virtual features. as well as The custom character is displayed in the user interface of the user device.

2. The computer-implemented method of claim 1, wherein the user interface includes one or more options for using the custom character, and further includes: The user device receives an option selection, the option corresponding to using the custom character to participate in a game session with the game name; Query the game manager for one or more character parameters associated with the game name; and Verify whether the custom character includes the character parameters associated with the game name, wherein the game session is initiated based on the verification.

3. The computer-implemented method as described in claim 2, further comprising: It has been preliminarily determined that the custom character does not include the character parameters associated with the game name; The machine learning model is used to generate one or more character modifications, wherein the one or more character modifications conform to the character parameters; as well as The character modification is applied to the custom character, wherein the custom character is verified based on the character modification.

4. The computer-implemented method of claim 1, further comprising using the machine learning model to generate one or more custom interactive storylines including the custom character, wherein the custom interactive storylines are presented in the virtual environment.

5. The computer-implemented method of claim 4, wherein the one or more custom interactive storylines are generated based on feedback from the user device.

6. The computer-implemented method of claim 1 further includes verifying whether the one or more content features conform to a set of review parameters associated with the user.

7. The computer-implemented method of claim 1, wherein the custom role is one of a set of different custom roles generated for including the virtual features and presented in the user interface, and wherein the user interface also presents a prompt to select one of the set of different custom roles.

8. The computer-implemented method of claim 1, further comprising generating design prompts to present to the user device, wherein the chat-based communication is received in response to the design prompts.

9. The computer-implemented method of claim 1, wherein the machine learning model is trained using historical game data about stored characters, including custom characters.

10. The computer-implemented method of claim 9, wherein the machine learning model is further trained using associated data from at least one of one or more additional users, user characteristics, game names, or game features.

11. The computer-implemented method of claim 1, wherein the custom role includes predicted virtual features, and wherein the custom role includes predicted virtual features associated with one or more stored custom roles.

12. The computer-implemented method as described in claim 1, further comprising: Identify one or more user characteristics associated with the user; as well as Using the machine learning model, additional virtual features are generated based on the identified user characteristics, wherein the generation of the custom role for use in the virtual environment is also based on the identified user characteristics.

13. The computer-implemented method of claim 1, further comprising generating one or more prompts, the one or more prompts presenting one or more additional virtual feature options selectable to be applied to the custom character.

14. A computing device, comprising: A communication interface that communicates with a user's device via a communication network, wherein the communication interface receives chat-based communications from the user device; as well as Processor, the processor executing instructions stored in memory, wherein the processor executing the instructions to: Machine learning models are used to interpret the chat-based communication, wherein interpreting the chat-based communication includes recognizing the meaning of the content in the chat-based communication and predicting that the content in the chat-based communication indicates one or more virtual features; and A custom character for use in a virtual environment is generated based on the identified meaning and predicted virtual characteristics, wherein the custom character includes the predicted virtual characteristics, and the display of the custom character is presented in the user interface of the user device.

15. The computing device of claim 14, wherein the user interface includes one or more options for using the custom character, and wherein the processor executes further instructions to: The user device receives an option selection, the option corresponding to using the custom character to participate in a game session with the game name; Query the game manager for one or more character parameters associated with the game name; and Verify whether the custom character includes the character parameters associated with the game name, wherein the game session is initiated based on the verification.

16. The computing device of claim 15, wherein the processor executes further instructions to: It has been preliminarily determined that the custom character does not include the character parameters associated with the game name; The machine learning model is used to generate one or more character modifications, wherein the one or more character modifications conform to the character parameters; and The character modification is applied to the custom character, wherein the custom character is verified based on the character modification.

17. The computing device of claim 14, wherein the processor executes further instructions to generate one or more prompts for the user, wherein the prompts include one or more additional virtual features applicable to the custom role that the user may approve or reject.

18. A non-transitory computer-readable storage medium, the computer-readable storage medium comprising instructions that, when executed by a computer, cause the computer to: Receive chat-based communications from the user's device via a communication network; Using a machine learning model to interpret the chat-based communication, wherein interpreting the chat-based communication includes recognizing the meaning of the content in the chat-based communication and predicting that the content in the chat-based communication indicates one or more virtual features; A custom role for use in a virtual environment is generated based on the identified meaning and predicted virtual features, wherein the custom role includes the predicted virtual features; and The custom character is displayed in the user interface of the user device.

Images