Peer-reviewed articles, research posters, technical reports, and presentation slides in Reinforcement Learning, Bio-Inspired AI, and Autonomous Navigation.
While deep reinforcement learning has achieved superhuman performance in fully observable, zero-sum games, environments characterized by high stochasticity and complex state spaces remain a significant challenge. This research introduces 'RainbowZero,' a novel algorithmic framework that synergizes the forward-planning capabilities of search-based algorithms (such as MuZero and Muesli) with advanced value estimation techniques and world models (e.g., Dreamer and Rainbow). We utilize the board game Settlers of Catan as our primary testbed due to its complex mechanics, varying board layouts, and high variance. The RainbowZero architecture serves as a proxy for complex, real-world optimization problems, specifically targeting autonomous decision-making in industrial operations.
We introduce RatSLAM 2.0, a novel biologically-plausible SLAM algorithm modernizing the original RatSLAM by integrating a new neocortical-hippocampal-entorhinal model based on the VectorHaSH framework. Our research generalizes the VectorHaSH memory scaffold to 3D spatial dimensions and non-square modules, confirming high capacity scaling linearly with module count. We benchmark Hebbian learning variations against analytic and iterative pseudoinverse methods, demonstrating that iterative pseudoinverse learning achieves superior stability for long-term associative memory. Evaluated in the AnimalAI environment, RatSLAM 2.0 successfully performs SLAM and pose tracking during continuous motion.