Salesforce.com

About: Salesforce.com is a based out in . It is known for research contribution in the topics: User interface & Object (computer science). The organization has 2418 authors who have published 2775 publications receiving 63956 citations.

Provable Self-Play Algorithms for Competitive Reinforcement Learning

Abstract: Self-play, where the algorithm learns by playing against itself without requiring any direct supervision, has become the new weapon in modern Reinforcement Learning (RL) for achieving superhuman performance in practice. However, the majority of exisiting theory in reinforcement learning only applies to the setting where the agent plays against a fixed environment; it remains largely open whether self-play algorithms can be provably effective, especially when it is necessary to manage the exploration/exploitation tradeoff. We study self-play in competitive reinforcement learning under the setting of Markov games, a generalization of Markov decision processes to the two-player case. We introduce a self-play algorithm---Value Iteration with Upper/Lower Confidence Bound (VI-ULCB)---and show that it achieves regret $\tilde{\mathcal{O}}(\sqrt{T})$ after playing $T$ steps of the game, where the regret is measured by the agent's performance against a \emph{fully adversarial} opponent who can exploit the agent's strategy at \emph{any} step. We also introduce an explore-then-exploit style algorithm, which achieves a slightly worse regret of $\tilde{\mathcal{O}}(T^{2/3})$, but is guaranteed to run in polynomial time even in the worst case. To the best of our knowledge, our work presents the first line of provably sample-efficient self-play algorithms for competitive reinforcement learning.

System and method for controlling access to aspects of an electronic messaging campaign

Abstract: A system for controlling access within an enterprise to information associated with recipients of an electronic message campaign of the enterprise sent to a plurality of recipient devices wherein the enterprise includes hierarchically structured Business Units having an enterprise level Business Unit at the highest level and a plurality of second level Business Units and an enterprise system communicatively coupled to a network and including an enterprise level device communicatively coupled to a plurality of second level devices includes a server and an electronic message engine The server is configured to assign an enterprise account to the enterprise system and to allow the enterprise level device to communicate selected portions of the recipient list. The electronic message engine is configured to generate electronic messages within a message campaign for sending to recipients identified by each of the second level devices from the selected portions of the recipient list.

Systems and methods for enhancing user engagement

Abstract: Methods and systems are provided for enhancing user engagement in a social network computing environment by using affinity indices to track the interaction among posts, topics, users, groups, and experts, and populating displays on user, group, and topic pages using the affinity indices.

Method and system for managing recent data in a mobile device linked to an on-demand service

Abstract: Systems and methods for managing recent data items in a database. A method typically includes determining whether a data object managed by an on demand service is designated as able to be accessed by a user at a mobile device and storing locally at a mobile device a plurality of most recently used items viewed for a data object designated as able to be accessed by a user at a mobile device. The method also typically includes determining a single most recently used set from among the stored plurality of most recently used items viewed for at least one data object designated as able to be accessed by a user at a mobile device.

Distributing relevant information to users of an enterprise network

Abstract: Various implementations are directed to systems, apparatus, computer-implemented methods and storage media for identifying a target set of users of an enterprise network to which to distribute a communication of enterprise-related information. For example, when a communication system receives a request to distribute a communication, the communication system analyzes the communication to identify a set of enterprise users that are predicted to find the information in the communication relevant, and especially, relevant from the enterprise's perspective. For example, the communication system can include a machine learning system that can construct, update and maintain a machine learning model of induction. In some implementations, the machine learning system trains the machine learning model by identifying contextual features of previously distributed communications, user traits of recipients of the previously distributed communications, and actions or inactions that indicate whether the recipients found the information in the communications relevant.