Neural-Robots

Abstract

Computational neuroscience provides powerful methodologies to address one of the greatest scientific challenges of our time: understanding how the nervous system controls behavior. Models of the nervous system allow for exploring the roles of neural circuits in sensory processing, cognition, behavior generation, and motor control. However, understanding how the nervous system controls behavior is impossible without the consideration of the body it evolved to function within. Embodiment addresses this critical aspect by closing the perception-action loop through real-world physical interactions. Robotics and physics-simulators effectively help bridge this gap by providing physical embodiment to computational neural models, facilitating tangible interactions with the environment. Furthermore, biological neural circuits have evolved to develop robust solutions to challenges that closely parallel those in robotics, including motor learning, adaptation to novel settings, robust locomotion, and navigation in complex environments. Harnessing solutions refined over millions of years of evolution, thus, holds immense potential to inspire the next breakthroughs in robotics.

This workshop at IROS 2025 brings together roboticists leveraging computational neuroscience to address robotics challenges and neuroscientists using robots to study neural function. Unlike previous related events, this workshop introduces cutting-edge approaches in neuroAI, neuromorphic engineering, embodied neural computation, and bio-inspired robotics to enhance the real-world applicability and robustness of robotic systems. The workshop promotes interdisciplinary collaboration, advancing fundamental understanding in neuroscience while simultaneously fostering innovative, bio-inspired solutions to contemporary robotic challenges.