PhySA: Decoupling Aesthetic and Physical Rewards for Humanoid Locomotion
cs.RO
Submitted: 2026-11-05
Updated: 2026-11-05
Code: https://github.com/krafton-ai/Prompt2Policy
Terminology
Sources
- Reward Design with Language Models
- RL-VLM-F: Reinforcement Learning from Vision Language Foundation Model Feedback
- ERA: Transforming VLMs into Embodied Agents via Embodied Prior Learning and Online Reinforcement Learning
- Look Before You Leap: Unveiling the Power of GPT-4V in Robotic Vision-Language Planning
- OpenAI Gym
- Retargeting Matters: General Motion Retargeting for Humanoid Motion Tracking
- Proximal Policy Optimization Algorithms
- GPT-4o System Card
- LLaVA-NeXT-Interleave: Tackling Multi-image, Video, and 3D in Large Multimodal Models
- Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning
- Do As I Can, Not As I Say: Grounding Language in Robotic Affordances
- Adam: A Method for Stochastic Optimization
- High-Dimensional Continuous Control Using Generalized Advantage Estimation
- Ablation Studies in Artificial Neural Networks
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation