CrossBFM: Distilling a Shared Latent Behavior Space Across Humanoid Embodiments

2026-09-29

AI Summary

This study addresses the high training costs of behavior foundation models and the challenge of transferring latent spaces across diverse robot morphologies by proposing the CrossBFM framework. Treating the latent space as a transferable asset, this method constructs a parameter-free unified encoder via retargeting correspondences and integrates forward-backward representations with PPO reinforcement learning to efficiently distill a cross-morphology shared behavior space. Experimental results demonstrate successful transfer across three prompting modalities on three humanoid robots, retaining 89% of morphology-specific performance while enabling zero-shot generalization. This approach substantially reduces training overhead, and its effectiveness has been validated on physical robots.

Abstract

Behavior Foundation Models (BFMs) give humanoids a promptable policy over a latent behavior space, enabling one single vector to represent a motion to imitate, a pose to reach, or a reward to maximize. Forward-Backward representations successfully produce such spaces, but at the cost of hundreds of GPU-hours for a single robot. Moreover, when the training process is repeated for a second robot, it produces a second space unrelated to the first, resulting in embodiment-specific latents that do not unify or transfer. We address these problems with CrossBFM, treating the latent space as the transferable asset for various embodiments. As retargeting provides frame-level cross-embodiment correspondence, we propose a unified encoder architecture with no robot-specific parameters for distilling the behavior space to address all training embodiments simultaneously in less than a GPU-hour. Following this encoder, latent-conditioned trackers turn the distilled latent into whole-body control in a conventional PPO training manner in just 10 more GPU-hours. On three distilled humanoids, all three prompting modes transfer: motion tracking with latent-conditioned policy losing only $0.025$ rad to its joint-conditioned counterpart, smooth goal reaching between poses with no falls, and reward optimization for all $41$ reward prompts. Our experiments further reveal that 1) regressing the encoder on a quarter of the motion corpus costs only $5\%$ of tracking performance and 2) training the encoder on a subset of robots and evaluating on an unseen one recovers up to $89\%$ of the tracking performance of seen robots, demonstrating cross-embodiment generalization to morphologically similar robots. We also verify the pipeline on real robots across all three prompting modes and with flow-based generated latents. Project website: https://dotandung.github.io/crossbfm/

Problem

• Behavior Foundation Models
• latent behavior space
• cross-embodiment transfer
• humanoid robots

Innovation

• Behavior Foundation Model
• Cross-embodiment Transfer
• Latent Space Distillation
• Unified Encoder
• Humanoid Control