Intelligence Is Not Only Reasoning
Published:
Posts by Tags
AI
Cognitive Bandwidth in the AI Agent Era
Published:
AI Agents
Intelligence Is Not Only Reasoning
Published:
Cognitive Bandwidth in the AI Agent Era
Published:
CoRL
[Paper Notes] DayDreamer: World Models for Physical Robot Learning - CoRL 2022
Published:
Key information
- This paper learns two models: a world model trained on off-policy sequences through supervised learning, and an actor-critic model to learn behaviors from trajectories predicted by the learned model.
- The data collection and learning updates are decoupled, enabling fast training without waiting for the environment. A learner thread continuously trains the world model and actor-critic behavior, while an actor thread in parallel computes actions for environment interaction.
Cognition
Cognitive Bandwidth in the AI Agent Era
Published:
Communication
Intelligence Is Not Only Reasoning
Published:
Dexterous Manipulation
Dynamics Models
Egocentric Data
Embodied AI
[Paper Notes] RLinf → RLinf-USER → RL-Co: A Full-Stack RL Pipeline for Embodied VLA Training
Published:
Human-to-Robot Transfer
Humanoid
Imitation Learning
Information Retrieval
Intelligence Is Not Only Reasoning
Published:
Intelligence
Intelligence Is Not Only Reasoning
Published:
Manipulation
Model-Based Control
Paper Notes
[Paper Notes] RLinf → RLinf-USER → RL-Co: A Full-Stack RL Pipeline for Embodied VLA Training
Published:
Personal Thoughts
Intelligence Is Not Only Reasoning
Published:
Cognitive Bandwidth in the AI Agent Era
Published:
Productivity
Cognitive Bandwidth in the AI Agent Era
Published:
RSS
[Paper Notes] SAM-RL: Sensing-Aware Model-Based Reinforcement Learning via Differentiable Physics-Based Simulation and Rendering - RSS 2023
Published:
Key information
- Model-based reinforcement learning (MBRL) is potentially more sample-efficient than model-free RL.
- Integration of differentiable physics-based simulation and rendering facilitates the learning process.
- Pipeline: Real2Sim -> Learn@Sim -> Sim2Real to produce efficient policies.
Reinforcement Learning
[Paper Notes] RLinf → RLinf-USER → RL-Co: A Full-Stack RL Pipeline for Embodied VLA Training
Published:
[Paper Notes] SAM-RL: Sensing-Aware Model-Based Reinforcement Learning via Differentiable Physics-Based Simulation and Rendering - RSS 2023
Published:
Key information
- Model-based reinforcement learning (MBRL) is potentially more sample-efficient than model-free RL.
- Integration of differentiable physics-based simulation and rendering facilitates the learning process.
- Pipeline: Real2Sim -> Learn@Sim -> Sim2Real to produce efficient policies.
[Paper Notes] DayDreamer: World Models for Physical Robot Learning - CoRL 2022
Published:
Key information
- This paper learns two models: a world model trained on off-policy sequences through supervised learning, and an actor-critic model to learn behaviors from trajectories predicted by the learned model.
- The data collection and learning updates are decoupled, enabling fast training without waiting for the environment. A learner thread continuously trains the world model and actor-critic behavior, while an actor thread in parallel computes actions for environment interaction.
Robotics
[Paper Notes] RLinf → RLinf-USER → RL-Co: A Full-Stack RL Pipeline for Embodied VLA Training
Published:
[Paper Notes] SAM-RL: Sensing-Aware Model-Based Reinforcement Learning via Differentiable Physics-Based Simulation and Rendering - RSS 2023
Published:
Key information
- Model-based reinforcement learning (MBRL) is potentially more sample-efficient than model-free RL.
- Integration of differentiable physics-based simulation and rendering facilitates the learning process.
- Pipeline: Real2Sim -> Learn@Sim -> Sim2Real to produce efficient policies.
[Paper Notes] DayDreamer: World Models for Physical Robot Learning - CoRL 2022
Published:
Key information
- This paper learns two models: a world model trained on off-policy sequences through supervised learning, and an actor-critic model to learn behaviors from trajectories predicted by the learned model.
- The data collection and learning updates are decoupled, enabling fast training without waiting for the environment. A learner thread continuously trains the world model and actor-critic behavior, while an actor thread in parallel computes actions for environment interaction.
Scaling Laws
Sim2Real
Systems
[Paper Notes] RLinf → RLinf-USER → RL-Co: A Full-Stack RL Pipeline for Embodied VLA Training
Published:
Unsupervised RL
VLA
[Paper Notes] RLinf → RLinf-USER → RL-Co: A Full-Stack RL Pipeline for Embodied VLA Training
Published:
Vision-Language-Action
World Models
update
Updating website
Published:
I just completed a major update to my personal website.
website
Updating website
Published:
I just completed a major update to my personal website.