Matthew Robards
Papers
1
Total Citations
6
H-Index
1
About
Matthew Robards is a researcher in reinforcement learning and machine learning, with a particular focus on kernel-based methods and temporal-difference learning. His work has contributed to the development of efficient algorithms for learning in high-dimensional state spaces, notably through the introduction of sparsity in kernel-based reinforcement learning. His most-cited paper, "Sparse Kernel-SARSA(λ) with an Eligibility Trace" (2011), has garnered 6 citations and presents a novel approach to combining kernel methods with eligibility traces to improve sample efficiency and computational tractability in online learning. This work addresses a key challenge in reinforcement learning: balancing the expressiveness of function approximation with the need for scalable, real-time updates. Robards’ contributions are particularly relevant for applications in robotics, autonomous systems, and any domain requiring adaptive decision-making under uncertainty. His research stands out for its rigorous mathematical foundation and practical algorithmic innovations, offering a bridge between theoretical advances and real-world deployment.
Research Focus
Key Achievements
Top Papers
- 1Sparse Kernel-SARSA(λ) with an Eligibility Trace6 citations · 2011