Lab Logo Background

Improbable AI Research Lab

Advancing the frontiers of robotics and artificial intelligence through innovative research in locomotion, manipulation, and machine learning. As a lab at MIT CSAIL, we develop cutting-edge technologies that push the boundaries of what robots and AI can achieve in complex, real-world environments.

Latest News

Recent updates from the lab

Jan 15 Prof. Agrawal receives the IIT Kanpur Young Alumnus Award
Jan 12 Prof. Agrawal receives 2024 IEEE Early Academic Career Award in Robotics and Automation

Whole-Body Locomotion

Developing advanced locomotion systems that enable robots to navigate complex, unstructured environments with human-like agility and efficiency. Our research spans legged robotics, dynamic walking, and adaptive locomotion strategies.

  • Legged robotics and dynamic walking
  • Terrain adaptation and navigation
  • Energy-efficient locomotion
  • Multi-modal mobility systems
Explore Research

Dexterous Manipulation

Pioneering dexterous manipulation capabilities that enable robots to interact with objects and environments with human-like precision and adaptability. Our work encompasses hardware design, control systems, and learning algorithms.

  • Dexterous hand design and control
  • Tactile sensing and feedback
  • In-hand manipulation
  • Human-robot interaction
Explore Research

General Machine Learning

Advancing fundamental machine learning algorithms and methodologies that drive innovation across robotics and AI. Our research focuses on developing robust, efficient, and interpretable learning systems.

  • Reinforcement learning algorithms
  • Multi-modal learning
  • Transfer learning and adaptation
  • Robust and interpretable AI
Explore Research