Valve Corporation

Anthony John Cox

John McDonald

Christian Carollo

Taylor Karsten Sherman

Martin Otten

Donald Ichiro Lambe

Alden Kroll

Adil Sardar

Richard Kaethler

Matt Rhoten

Brian R Levinthal

Mark Zbikowski

Gabe Van Engel

Gabriel Van Engel

Described herein are, among other things, techniques, devices, and systems for generating one or more trained machine-learning models used for generating content-item recommendations. Also described herein are techniques, devices, and systems for applying a consumption history of a particular user to the trained model(s) to generate score data indicating a correlation between each content-item title and the consumption history, as well as modifying this score data using one or more biasing factors for generating result data. In addition, the techniques, devices, and systems may use this result data, along with received user input, for determining an order in which to present one or more content items to the user. For example, this may include determining which content items to recommend to a user and in which order to do so.

Content-item recommendations

Described herein are, among other things, techniques, devices, and systems for generating one or more trained machine-learning models. Also described herein are techniques, devices, and systems for applying a consumption history of a particular user to the trained model(s) to generate score data indicating a correlation between each content-item title and the consumption history. The techniques then determine a ranked list of content items having a highest correlation to the consumption history, which may be used to retrieve videos associated with the most-correlated content items for generating a compilation video composed of these retrieved videos.

Custom compilation videos

A trained machine learning model(s) is used to determine scores (e.g., trust scores) for user accounts registered with a video game service, and the scores are used to match players together in multiplayer video game settings. In an example process, a computing system may access data associated with registered user accounts, provide the data as input to the trained machine learning model(s), and the trained machine learning model(s) generates the scores as output, which relate to probabilities of players behaving, or not behaving, in accordance with a particular behavior while playing a video game in multiplayer mode. Thereafter, subsets of logged-in user accounts executing a video game can be assigned to different matches based at least in part on the scores determined for those logged-in user accounts, and the video game is executed in the assigned match for each logged-in user account.

Machine-learned trust scoring for player matchmaking

Techniques are described for automatically reducing cheating in an interactive execution environment, such as to perform automated operations to detect and stop use of cheat software in an online game environment, and to restrict subsequent access to the online game environment for users who are identified as using cheat software. The techniques may include using deep learning techniques to train one or more models to classify particular types of gameplay actions as being unauthorized if cheat software use is detected, including to determine a likelihood of whether a separate cheat detection decision system would decide that particular gameplay actions are authorized or not authorized if the additional cheat detection decision system assesses those gameplay actions, and then using the trained model(s) and in some cases the additional cheat detection decision system for the cheat software detection and prevention.

Automatically reducing use of cheat software in an online game environment

Growing community of inventors

Seattle, WA, United States of America

Anthony John Cox