Sony

Sony's 2 AI and machine learning patents both leverage specialized models to automate tasks traditionally requiring human judgment or manual configuration. The first analyzes sports gameplay by tracking objects and computing what Sony calls a "point excitement matrix," which scores moments based on emotional intensity rather than just detecting basic events like goals or hits. The second uses multi-modal feature extraction to translate in-game happenings into smart home responses, learning from user feedback to determine which gameplay states should trigger which environmental changes (like dimming lights during stealth sequences or adjusting temperature during tense moments), with a latency compensation mechanism that predicts future game states to trigger device changes before the moment actually occurs.

The 3 cloud gaming patents tackle different aspects of Streaming playable content over variable network conditions. One decouples text rendering from scene rendering at the draw call level, transmitting UI elements and captions separately so they remain crisp even when bandwidth constraints force the visual scene into heavy compression. Another combines neural network-based latency prediction with client-side AI upscaling, allowing the system to proactively downgrade certain stream elements before connection quality degrades while the client reconstructs detail locally. The third converts recorded gameplay segments into playable mini-game slices that others can experience interactively rather than watching passively, attaching frame-level social tags to specific moments that persist with the content as it gets shared.