We open-source many of our research projects. See the publications page for a comprehensive list of our papers along with released software.

Reinforcement Learning

• Guided Policy Search, a library for model-based deep reinforcement learning.
• Soft Q-learning, a library for model-free maximum entropy reinforcement learning.
• Soft Actor-Critic, an off-policy actor-critic framework for model-free maximum entropy reinforcement learning.
• Inverse RL, a library for model-free inverse reinforcement learning.
• Generalized Computation Graphs, a library for deep reinforcement learning.
• Divide and Conquer Reinforcement Learning, a library for model-free deep reinforcement learning.
• RL Kit, a reinforcement learning framework and algorithms implemented in PyTorch.

Computer Vision

• video_prediction, a Tensorflow model for deep video prediction.
• objectattention, a library for object detection for robotics.
• revolver, a library for few-shot segmentation using PyTorch.