This month's AI & Machine Learning category includes 7 filed patent applications from 4 companies: Sony (3), Skydance Silicon Valley (2), Nvidia (1), and WMC Analytics (1).
The patents focus on machine learning systems that adapt gaming experiences across different contexts and platforms. Sony's applications cover cross-platform frame adaptation for mobile streaming, trophy prediction from player data analysis, and personalized trailer generation based on console history. Skydance Silicon Valley filed applications for automated camera control systems that use neural networks to select optimal viewpoints and manage cinematic cuts, while Nvidia's application describes natural language driven shader customization for game streaming and WMC Analytics filed technology for AI-powered gameplay analysis and skill gap identification.
Sony received 3 patents covering AI-driven personalization and cross-platform optimization. The first application describes a system that understands which game elements matter most for gameplay and selectively enhances their visibility when streaming from console to mobile, rather than simply resizing the entire frame. Another filed application analyzes aggregated gameplay data from millions of players to identify which actions correlate most strongly with unlocking specific trophies or rare items, providing quantified likelihood scores instead of relying on manual guides. The third patent covers a system that builds individualized game trailers by selecting and combining video segments based on a player's console usage history and behavioral patterns.
Skydance Silicon Valley filed 2 patents focused on automated camera systems for real-time games. One application describes a neural network that continuously scores available camera angles and selects the optimal viewpoint without relying on predefined trigger zones or hardcoded rules. The other covers a system that predicts how long each player needs to reorient after a camera cut and allows subtle player input to influence automated camera selection without requiring direct manual control, maintaining cinematic flow while adapting to individual comfort levels.
WMC Analytics received 1 patent for a replay analysis system that parses game files, classifies in-game events, and generates personalized training recommendations. The system identifies which skill would yield the greatest improvement relative to a player's goals rather than simply flagging the weakest absolute skill, focusing coaching effort on the highest-leverage areas for development.
Nvidia filed 1 patent covering AI-generated shader code for game streaming. The system translates natural language descriptions into executable shader programs, allowing players to customize visual effects by describing what they want in plain English rather than selecting from a predetermined menu of options.
All data sourced from USPTO patent filings. Google Patents may take several weeks to index recent publications. If a link is unavailable, search for the patent number at USPTO Patent Public Search.