The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Models for Human-Robot Collaboration with Julie Shah - #538
7 snips
Nov 22, 2021 Julie Shah, a professor at MIT, specializes in interactive robotics. In this engaging discussion, she explains how robots can predict human behavior, vital for effective collaboration. Shah highlights her ambitious vision for a field robot that operates independently of human control. She also delves into the nuances of cross-training robots and humans for improved teamwork and the challenges of adapting to human unpredictability in dynamic environments. The conversation is a captivating glimpse into the future of human-robot partnerships.
AI Snips
Chapters
Transcript
Episode notes
Path to Robotics
- Julie Shah's interest in aerospace stemmed from a childhood fascination with space and a book at space camp.
- This led her to MIT, where she specialized in control theory and human factors engineering.
Effective Teamwork
- Effective human team members anticipate their partners' actions and adapt their plans.
- Shah's lab focuses on building AI models that emulate these human capabilities.
AI Modeling Human Thought
- Inferring human cognitive states for AI requires well-defined problems and context-specific considerations.
- Unsupervised learning is unlikely to succeed; eliciting relevant inputs from humans is key.