I-BFM: Reward-Conditioned Robust Humanoid Interaction via Unsupervised Reinforcement Learning

arXiv:2610.06129 · cs.RO · Submitted 2026-10-05 · Read on arXiv

cs.RO

Submitted: 2026-10-05

Updated: 2026-10-05

Comments: 9 pages, including appendix

Project page: https://iamhardworking.github.io/I-BFM/Reward

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Reinforcement Learning — 1× in this paper · explained in 4 episodes

Sources

• BFM-Zero: A Promptable Behavioral Foundation Model for Humanoid Control Using Unsupervised Reinforcement Learning
• SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control
• BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusion
• HDMI: Learning Interactive Humanoid Whole-Body Control from Human Videos
• OmniRetarget: Interaction-Preserving Data Generation for Humanoid Whole-Body Loco-Manipulation and Scene Interaction
• HumanX: Toward Agile and Generalizable Humanoid Interaction Skills from Human Videos
• OmniContact: Chaining Meta-Skills via Contact Flow for Generalizable Humanoid Loco-Manipulation
• Zero-Shot Whole-Body Humanoid Control via Behavioral Foundation Models
• Behavior Foundation Model for Humanoid Robots
• FALCON: Learning Force-Adaptive Humanoid Loco-Manipulation
• TWIST: Teleoperated Whole-Body Imitation System
• VIRAL: Visual Sim-to-Real at Scale for Humanoid Loco-Manipulation
• Opening the Sim-to-Real Door for Humanoid Pixel-to-Action Policy Transfer
• HITTER: A HumanoId Table TEnnis Robot via Hierarchical Planning and Learning
• Learning Athletic Humanoid Tennis Skills from Imperfect Human Motion Data
• PhysHSI: Towards a Real-World Generalizable and Natural Humanoid-Scene Interaction System
• LessMimic: Long-Horizon Humanoid Interaction with Unified Distance Field Representations
• Learning One Representation to Optimize All Rewards
• INTACT: Isomorphic Intent-to-Action Learning for Search-Free World Models

Related papers

• FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
• MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
• RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
• ARC: A Reasoning Recipe for Robot Foundation Models
• HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
• APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies