Max Robotics — Global Service Robot Directory + US Certification

About

Haruki Nishimura is a rising researcher at the forefront of safe learning-based control for robotics, with a focus on ensuring that autonomous systems operate reliably even when faced with unfamiliar scenarios. His work centers on the critical challenge of bridging the gap between high-performance imitation learning policies and real-world safety, particularly in manipulation and interactive planning. Nishimura’s major contributions include pioneering uncertainty-aware methods for runtime failure detection, allowing robots to identify when they are operating outside their training distribution without requiring explicit failure data. He also introduced In-Distribution Barrier Functions, a self-supervised filter that prevents learned controllers from encountering out-of-distribution states, directly addressing a core vulnerability in modern robotic systems. His research on risk-aware prediction for robust planning further advances the field by mitigating the underestimation of long-tail safety-critical events in interactive environments. With each of his most-cited papers garnering 2–5 citations in top venues, Nishimura’s work is gaining traction for its practical approach to deploying complex learned policies safely. His comparative studies on evolutionary methods and reinforcement learning for virtual creatures also showcase a broader interest in scalable learning paradigms. Nishimura’s contributions are essential reading for students and researchers aiming to build trustworthy, high-performance autonomous systems.