Dorsa Sadigh "Safe And Adaptive Human-Robot And Multi-agent Interactions"

Advancing human-robot interaction requires shifting from rigid, abstract mathematical models to predictive behavior models that capture human intent and adaptation. Current robotics research often relies on one-shot interactions, yet creating robots for home environments necessitates a focus on long-term, repeated interactions where both the human and the robot evolve. Dorsa, an assistant professor at Stanford, emphasizes that effective coordination—such as in assistive feeding—demands a holistic approach that integrates algorithmic intelligence with physical design. By moving beyond traditional disciplinary boundaries and incorporating diverse modalities like language, gaze, and gestures, researchers can develop systems that are not only safe but also capable of meaningful collaboration. This interdisciplinary perspective, combined with a willingness to challenge established research norms, remains essential for bridging the gap between laboratory prototypes and real-world, commercial applications.