Shotaro Miwa

About

Shotaro Miwa is a researcher at the forefront of making intelligent systems more transparent and resilient. His work primarily addresses two critical challenges in modern artificial intelligence: explainability in deep reinforcement learning (DRL) and hardware fault tolerance in autonomous machines. In his highly cited 2024 paper, Miwa pioneered the application of Layer-wise Relevance Propagation (LRP) to robotic domains, enabling researchers to visualize and understand the decision-making processes of DRL agents—a breakthrough that directly tackles the "black box" problem in robotics. This work has already garnered 9 citations, underscoring its significance for the AI safety community. Concurrently, Miwa is advancing industrial automation by enhancing hardware fault tolerance through reinforcement learning policy gradient algorithms. His innovative approach moves beyond traditional hardware duplication, allowing machines to dynamically adapt to faults using learned policies. With a clear focus on bridging the gap between high-performance AI and practical, trustworthy deployment, Miwa’s contributions are shaping a future where autonomous systems are not only powerful but also interpretable and robust against real-world failures.

Research Focus

Key Achievements

Top Papers

1. 1
   Explainability of deep reinforcement learning algorithms in robotic domains by using Layer-wise Relevance Propagation

   9 citations · 2024
2. 2
   Enhancing Hardware Fault Tolerance in Machines with Reinforcement Learning Policy Gradient Algorithms

   2 citations · 2024

Key Collaborators

Contact & Links