Microsoft Technology Licensing, LLC

Brent Jaron Hecht

Jaime Brooks Teevan

Longqi Yang

Mengting Wan

Meredith June Morris

Daniel John Liebling

Cao Lu

Jing Ma

Kiran Tomlinson

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. The method also includes calculating a transferability metric for each per-organization recommendation model based on results obtained during validation, determining a specified number of organizations including the best-transferring per-organization recommendation models based on the calculated transferability metrics, and training an inductive multi-organization recommendation model using the enterprise application data from the specified number of organizations. The method further includes utilizing the trained inductive multi-organization recommendation model to provide user recommendations to the remote organization computing systems during execution of the enterprise application.

Targeted training of inductive multi-organization recommendation models for enterprise applications

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. The method further includes estimating the ITE for each node under the treatment assignment based on the learned confounder representation and the modeled high-order interference representation for each node.

Individual treatment effect estimation under high-order interference in hypergraphs

A system is described in which a user can add one or more automated agents as 'friends' in a social network service. In operation, an automated agent observes an information need expressed by the user via the social network service, e.g., in the form of an original message posted to the social network service; determines whether it is appropriate to reply to the information need; uses automated functionality to generate a reply message to the information need (if it is deemed appropriate to reply to the information need); and sends the reply message to a target destination, such as a social network page associated with the user. For example, without limitation, one type of automated agent performs a question-answering function. Another type of automated agent performs a social referral service.

Designating automated agents as friends in a social network service

Growing community of inventors

Redmond, WA, United States of America

Brent Jaron Hecht