Abstract: The present technology provides a smart notification service that is machine-learning driven and integrates with smart channels for determining how the user is receiving notifications and smart notification generators for determining what notifications should be sent and to whom, based on what notifications have been sent previously to the users and how they interacted with the previously sent notifications. The machine-learning model is trained by previously sent notifications to optimize the kind of notifications being sent, who receives the notifications, and in what cadence they receive such notifications.

Abstract: The present technology provides native access control handling based on curated data received through third-party authentication protocols. Authentication of a user account is based on a set of curated data received from a third-party application with respect to a third-party application user account shared by a same user. The user account is a member of a server that supports near real-time communications through a client application associated with the user account. The set of curated data is received in response to receiving authentication based on a third-party authentication protocol associated with the third-party application. The curated set of data can be displayed in a profile of the user account as well as for providing an access privilege to the user account in the server.

Abstract: The present technology provides an emoji recommendation service that measures a similarity score between a global emoji embedding space of an online community with a user embedding space generated from the global emoji embedding space and generates a ranking of top emojis based on the measured similarity score. The emoji recommendation service further measures an averaged ranking of top emojis based on the similarity score and term frequency scores in the online community. The emoji recommendation service trains a machine learning algorithm to arrange vectors representing emojis in the global emoji embedding space.

Abstract: The disclosed technology addresses the need in the art to enable users to broadcast a status that indicates that a user is inviting others to join their livestream, such that they are either prepared to livestream or is livestreaming. An invite-to-join status may be limited to the specific friends and community members of the user's choosing. If the user has no viewer yet, the user will not broadcast a livestream and will continue with what they are doing, such as playing a game or watching a movie. Once a first viewer joins, then the user screen may be broadcasted into a livestream.

Abstract: The present technology extends to methods, systems, and computer program products for expanding semantic classes via user feedback. Aspects of the technology learn how a set of labels can be expanded from user-generated tags. Text labels applied by human reviewers to digital content can be inspected and compared to one another. When a threshold of human-generated text tags contain similar terminology, the set of labels can be expanded to define a representation of the similar terminology. Similar terminology can include terms that originate from the same base term, are synonyms, are more specific terms related to a general term category, etc. Similar terminology can be consolidated into a defining term that is used to generate a new (more granular) label or a new top level label. Accordingly, new semantic classes can be discovered from user-generated feedback. New semantic classes can provide a more granular representation of content item classification.

Abstract: The present invention extends to methods, systems, and computer program products for detecting online contextual evolution of linguistic terms. Within messaging sources, some users may actively attempt to (relatively quickly) shift the meaning of a word or term. Some users may attempt to perjorate a word or term to have a more toxic meaning. Other users may attempt to reappropriate a word or term to have a less toxic or even a positive meaning. Aspects of the invention identify shifts in implied meanings of words and/or phrases over time. As such, emerging forms of harassment can be identified more quickly. Aspects of the invention can utilize users' behavioral histories as well as messaging structures to improve confidence when identifying term evolution. Machine learning algorithms can be configured to identify term evolution reducing workload on human moderators.

Abstract: The disclosed technology addresses the need in the art for to facilitate user accounts participating in live audio-video sessions such as voice and/or video calls, or watching game play remotely to capture a clip of the video call or game play after seeing some event during the live audio-video session. The present technology can maintain a rolling buffer of the audio-video streams such that a user account can retroactively capture a clip of the audio-video streams after it has already been presented.

Abstract: The present technology utilizes data about diverse types of entities to inform machine learning models. The present technology can ingest social network data that includes nodes from diverse entity types, and utilize information implicit in relationships between one entity type with another entity type. The present technology creates a plurality of embedding spaces for representing entities of different types as vectors. The vectors identify a first entity, a second entity, and a relationship between the entities. The data from the embedding spaces can be input into a model configured to classify entities as likely to be associated with an explored characteristic, and output of a classification of entities that are likely associated with the explored characteristic. The model is configured to relate the different types of entities to classify entities as likely to be associated with the explored characteristic.

Abstract: The present technology utilizes data about diverse types of entities to inform machine learning models. The present technology can ingest social network data that includes nodes from diverse entity types, and utilize information implicit in relationships between one entity type with another entity type. The present technology creates a plurality of embedding spaces for representing entities of different types as vectors. The vectors identify a first entity, a second entity, and a relationship between the entities. The data from the embedding spaces can be input into a model configured to classify entities as likely to be associated with an explored characteristic, and output of a classification of entities that are likely associated with the explored characteristic. The model is configured to relate the different types of entities to classify entities as likely to be associated with the explored characteristic.

Abstract: The present technology provides real-time message moderation that checks messages against a list of trigger checks and determines whether or not the message should be blocked before the message is sent to other members of the community. The real-time message moderation uses a rules system that enables administrators to generate a customized sets of rules, such as custom keyword filter that blocks a message if it contains a word that matches a keyword associated with the custom keyword filter rule, and wherein custom keywords may be added by a moderator of the server. Moderators can report issues with past blocked messages to train a machine-learning model about a bad flag.

Abstract: The present technology provides real-time message moderation that checks messages against a list of trigger checks and determines whether or not the message should be blocked before the message is sent to other members of the community. The real-time message moderation uses a rules system that enables administrators to generate a customized sets of rules, such as custom keyword filter that blocks a message if it contains a word that matches a keyword associated with the custom keyword filter rule, and wherein custom keywords may be added by a moderator of the server. Moderators can report issues with past blocked messages to train a machine-learning model about a bad flag.

Abstract: The present invention extends to methods, systems, and computer program products for detecting online contextual evolution of linguistic terms. Within messaging sources, some users may actively attempt to (relatively quickly) shift the meaning of a word or term. Some users may attempt to perjorate a word or term to have a more toxic meaning. Other users may attempt to reappropriate a word or term to have a less toxic or even a positive meaning. Aspects of the invention identify shifts in implied meanings of words and/or phrases over time. As such, emerging forms of harassment can be identified more quickly. Aspects of the invention can utilize users' behavioral histories as well as messaging structures to improve confidence when identifying term evolution. Machine learning algorithms can be configured to identify term evolution reducing workload on human moderators.

Abstract: The present technology extends to methods, systems, and computer program products for expanding semantic classes via user feedback. Aspects of the technology learn how a set of labels can be expanded from user-generated tags. Text labels applied by human reviewers to digital content can be inspected and compared to one another. When a threshold of human-generated text tags contain similar terminology, the set of labels can be expanded to define a representation of the similar terminology. Similar terminology can include terms that originate from the same base term, are synonyms, are more specific terms related to a general term category, etc. Similar terminology can be consolidated into a defining term that is used to generate a new (more granular) label or a new top level label. Accordingly, new semantic classes can be discovered from user-generated feedback. New semantic classes can provide a more granular representation of content item classification.

Abstract: The present invention extends to methods, systems, and computer program products for detecting online contextual evolution of linguistic terms. Within messaging sources, some users may actively attempt to (relatively quickly) shift the meaning of a word or term. Some users may attempt to perjorate a word or term to have a more toxic meaning. Other users may attempt to reappropriate a word or term to have a less toxic or even a positive meaning. Aspects of the invention identify shifts in implied meanings of words and/or phrases over time. As such, emerging forms of harassment can be identified more quickly. Aspects of the invention can utilize users' behavioral histories as well as messaging structures to improve confidence when identifying term evolution. Machine learning algorithms can be configured to identify term evolution reducing workload on human moderators.

Abstract: The present invention extends to methods, systems, and computer program products for expanding semantic classes via user feedback. Aspects of the invention learn how a set of labels can be expanded from user-generated tags. Text labels applied by human reviewers to digital content can be inspected and compared to one another. When a threshold of human-generated text tags contain similar terminology, the set of labels can be expanded to define a representation of the similar terminology. Similar terminology can include terms that originate from the same base term, are synonyms, are more specific terms related to a general term category, etc. Similar terminology can be consolidated into a defining term that is used to generate a new (more granular) label or a new top level label. Accordingly, new semantic classes can be discovered from user-generated feedback. New semantic classes can provide a more granular representation of content item classification.