Adeia filed 1 patent in the AI & Machine Learning category.
The filing covers a time-aware deep learning framework designed for gaming applications, where AI opponents can adapt gameplay to fit within user-defined time windows while maintaining consistent difficulty levels. This approach addresses resource allocation challenges that arise when managing AI-controlled opponents in games with temporal constraints.
The single AI and machine learning patent tackles a persistent problem in games with AI opponents: fitting a complete gaming session into whatever time the player has available without making the experience feel rushed, padded, or inconsistent. Traditional approaches either swap between different AI models mid-session or adjust difficulty on the fly, both of which create jarring player experiences and burn through computing resources inefficiently. This framework embeds the time constraint directly into the Monte Carlo Tree Search algorithm itself, using a modified upper confidence bound calculation alongside dual neural networks that balance two competing goals at once. One network optimizes for winning or losing at an appropriate rate while the other ensures the session concludes within the target duration, preventing the common scenario where difficulty spikes awkwardly near the end or the AI clearly stalls to fill time.
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.