Abstract: A system and method and for optimizing cross-team information flow in a communication network includes receiving, from a communication application, via a network, a plurality of candidate post items for display to a first user of an organization, each candidate post item being a post item published by another user of the organization and being a post that is accessible to the first user. A communication knowledge network graph which represents communication events that have occurred between users of the organization is then generated where each communication event is represented by a first node that represents a sender, a second node that represents a receiver and an edge that represents the communication event from the sender to the receiver.

Abstract: A data processing system implements a framework that utilizes in-situ user interactions as a source of feedback for improving the training of LLMs to generate outputs that align with user preferences. The framework includes a user preference evaluation pipeline analyzes in-situ user interactions with the LLMs and generates preference information that can be used to improve the training of the LLM to improve the alignment of the LLM with user preferences. The user preference evaluation pipeline includes a feedback signal identification unit that identifies explicit and/or implicit feedback provided by the users in response to content output by the LLM in response to a user prompt. The feedback signal identification unit estimates user satisfaction with a set of satisfaction rubrics and user dissatisfaction with a set of user dissatisfaction rubrics to generate user preference data that can be used to align an LLM with these user preferences.

Abstract: Various embodiments discussed herein relate to prompting a model, such as a Large Language Model (LLM), to ingest natural language clustering instructions and generate corresponding natural language clustering information, such as a cluster description and/or a cluster label without the need to generate any numeric text embeddings.

Abstract: Various embodiments discussed herein relate to prompting a model, such as a Large Language Model (LLM), to ingest natural language clustering instructions and generate corresponding natural language clustering information, such as a cluster description and/or a cluster label without the need to generate any numeric text embeddings.

Abstract: Personally-stylized content can be generated without fine tuning a model. A personally-stylized content generation method can include receiving a first request for first content to be stylized in a style of written prose previously produced by a user, applying a previously trained retriever model to the first request to obtain second content previously produced by the user resulting in obtained content, populating a prompt with the obtained content and the first request resulting in an augmented prompt, providing the augmented prompt to a large language model (LLM), receiving personally-stylized content from the LLM, the personally-stylized content including elements of the style of the written prose of the user, and providing the personally-stylized content to the user.

Abstract: Various embodiments discussed herein relate to prompting a model, such as a Large Language Model (LLM), to ingest natural language clustering instructions and generate corresponding natural language clustering information, such as a cluster description and/or a cluster label without the need to generate any numeric text embeddings.

Abstract: Systems and methods for open-domain dialogue segmentation and state tracking are provided. In particular, a computing device may obtain and analyze a dialogue in near real-time, generate a structured prompt template for a state prediction model based on the dialogue, and generate a structured output using the state prediction model based on the structured prompt template. The structured output includes a turn summary and state labels for each dialogue turn.

Abstract: Methods, computer systems, computer-storage media, and graphical user interfaces are provided for efficiently generating and using intent taxonomies. In embodiments, training data, including data requests for information, is obtained. Thereafter, a model prompt to be input into a large language model is generated. The model prompt includes an instruction to generate an intent taxonomy, an indication of the training data to use for generating the intent taxonomy, and a taxonomy attribute desired to be used as criteria to generate a quality intent taxonomy. An intent taxonomy that includes user intent classes is obtained as output from the large language model. The intent taxonomy is analyzed to determine whether the intent taxonomy is valid. When the intent taxonomy is determined as valid, the intent taxonomy is provided for use in identifying user intent, and when the intent taxonomy is determined as invalid, the intent taxonomy is refined.

Abstract: A system and method for automatically generating a workplan schedule for a team of employees in an organization includes receiving a request to create a workplan schedule for the team, the team including two or more team members who are employees of the organization and the workplan schedule identifying a work location for the team members. One or more collaborators for one or more of the team members is detected, where the collaborator is an employee of the organization who is not a member of the team and with whom at least one of the team members collaborate.

Abstract: A system and method and for optimizing cross-team information flow in a communication network includes receiving, from a communication application, via a network, a plurality of candidate post items for display to a first user of an organization, each candidate post item being a post item published by another user of the organization and being a post that is accessible to the first user. A communication knowledge network graph which represents communication events that have occurred between users of the organization is then generated where each communication event is represented by a first node that represents a sender, a second node that represents a receiver and an edge that represents the communication event from the sender to the receiver.

Abstract: An enterprise system server, a computer-readable storage medium, and a method for targeted training of inductive multi-organization recommendation models for enterprise applications are described herein. The method includes receiving enterprise application data from remote organization computing systems executing the enterprise application, training per-organization recommendation models for a subset of the organizations, and validating each per-organization recommendation model on enterprise application data corresponding to one or more other organizations.

Abstract: A computing system, computer-readable storage medium, and method for individual treatment effect (ITE) estimation under high-order interference in hypergraphs are described herein. The method includes accessing, via a processor, a hypergraph dataset including multi-way interactions among nodes within each hyperedge of a corresponding hypergraph, where the hypergraph dataset corresponds to a treatment assignment for each node. The method includes performing representation learning on the hypergraph dataset to control for confounders corresponding to features of each node and to learn a confounder representation for each node. The method also includes modeling a high-order interference representation for each node by propagating the learned confounder representation and the treatment assignment for each node through a hypergraph neural network.

Abstract: An enterprise system server, a computer-readable storage medium, and a method for targeted training of inductive multi-organization recommendation models for enterprise applications are described herein. The method includes receiving enterprise application data from remote organization computing systems executing the enterprise application, training per-organization recommendation models for a subset of the organizations, and validating each per-organization recommendation model on enterprise application data corresponding to one or more other organizations.

Abstract: The present concepts relate to a differentiable user-item co-clustering (“DUICC”) model for recommendation and co-clustering. Users' interaction with items (e.g., content) may be centered around information co-clusters—groups of items and users that exhibit common consumption behavior. The DUICC model may learn fine-grained co-cluster structures of items and users based on their interaction data. The DUICC model can then leverage the learned latent co-cluster structures to calculate preference stores of the items for a user. The top scoring items may be presented to the user as recommendations.

Abstract: Methods, computer systems, computer-storage media, and graphical user interfaces are provided for enhancing remote work productivity data. In one embodiment, worker data is obtained for a set of target workers and a set of control workers. The target workers are generally identified as working in the office prior to a treatment date and working remote after a treatment date, and the control workers are generally identified as working remote before and after the treatment date. Thereafter, an effect of working remotely is determined using a difference-in-differences model that measures differences in observational changes between the set of target workers and the set of control workers measured before and after the treatment date. The effect of working remotely can be used to generate a generating a productivity score indicating an extent of productivity in relation to the set of target workers.

Abstract: The present concepts relate to a differentiable user-item co-clustering (“DUICC”) model for recommendation and co-clustering. Users' interaction with items (e.g., content) may be centered around information co-clusters—groups of items and users that exhibit common consumption behavior. The DUICC model may learn fine-grained co-cluster structures of items and users based on their interaction data. The DUICC model can then leverage the learned latent co-cluster structures to calculate preference stores of the items for a user. The top scoring items may be presented to the user as recommendations.

Abstract: A digital medium environment includes an asset processing application that performs editing of assets. A projection function is trained using pairs of actions pertaining to software edits, and assets resulting from the actions to learn a joint embedding between the actions and the assets. The projection function is used in the asset processing application to recommend software actions to create an asset, and also to recommend assets to demonstrate the effects of software actions. Recommendations are based on ranking distance measures that measure distances between actions representations and asset representations in a vector space.

Abstract: This disclosure involves personalizing user experiences with electronic content based on application usage data. For example, a user representation model that facilitates content recommendations is iteratively trained with action histories from a content manipulation application. Each iteration involves selecting, from an action history for a particular user, an action sequence including a target action. An initial output is computed in each iteration by applying a probability function to the selected action sequence and a user representation vector for the particular user. The user representation vector is adjusted to maximize an output that is generated by applying the probability function to the action sequence and the user representation vector. This iterative training process generates a user representation model, which includes a set of adjusted user representation vectors, that facilitates content recommendations corresponding to users' usage pattern in the content manipulation application.

Abstract: This disclosure involves personalizing user experiences with electronic content based on application usage data. For example, a user representation model that facilitates content recommendations is iteratively trained with action histories from a content manipulation application. Each iteration involves selecting, from an action history for a particular user, an action sequence including a target action. An initial output is computed in each iteration by applying a probability function to the selected action sequence and a user representation vector for the particular user. The user representation vector is adjusted to maximize an output that is generated by applying the probability function to the action sequence and the user representation vector. This iterative training process generates a user representation model, which includes a set of adjusted user representation vectors, that facilitates content recommendations corresponding to users' usage pattern in the content manipulation application.

Abstract: A digital medium environment includes an asset processing application that performs editing of assets. A projection function is trained using pairs of actions pertaining to software edits, and assets resulting from the actions to learn a joint embedding between the actions and the assets. The projection function is used in the asset processing application to recommend software actions to create an asset, and also to recommend assets to demonstrate the effects of software actions. Recommendations are based on ranking distance measures that measure distances between actions representations and asset representations in a vector space.