Research output

Publications

Papers and ongoing research projects. Conference versions and preprints of the same work appear as a single entry.

Conference Papers

NeurIPS 20262026 · AcceptedCo-first author

InfCoiL: Coordinated Planner-Controller Learning for Closed-Loop Physics-Based Human-Object Interaction

Yude Zou*, Yifei Yao*, Yiwei Hao*, Hanqing Wang, Xing Gao

* Equal contribution (co-first authors).

Make generated motion executable. InfCoiL lets a flow-matching planner and a physics-based controller improve each other through shared rollout feedback, alternating their RL updates when progress plateaus.

Watch demos

Closed-loop table transport

Qualitative S=2 planner / C45k controller rollout from the project page. The goal ring turns green on stable arrival. Five-frame display smoothing and a final-pose hold are part of the source video; this clip is separate from the paper benchmark. Source

Replanning through tripod transport

A second object and transport task from the same qualitative release, with the same display smoothing and final-pose hold. Source
BibTeX

Official project citation

@inproceedings{zou2026infcoil,
  title     = {{InfCoiL}: Coordinated Planner-Controller Learning for Closed-Loop Physics-Based Human-Object Interaction},
  author    = {Zou, Yude and Yao, Yifei and Hao, Yiwei and Wang, Hanqing and Gao, Xing},
  booktitle = {Advances in Neural Information Processing Systems},
  year      = {2026}
}
ECCV 20262026Co-first author

Exploratory, Communicative, and Deployable: Vision-Driven Embodied Agents for Open-World Mobile Manipulation

Boyu Mi*, Mengchen Ma*, Yifei Yao*, Xing Gao, Junting Chen, Yangzi Li, Zihou Zhu, Guohao Li, Zhenfei Yin, Tai Wang, Yao Mu, Jiangmiao Pang, Hanqing Wang

* Equal contribution (co-first authors).

Vision-driven mobile manipulation through active exploration, visual grounding, and interactive intent disambiguation, with a shared interface for simulation and real robots.

BibTeX

Official project citation

@inproceedings{mi2026real,
  title = {Exploratory, Communicative, and Deployable: Vision-Driven Embodied Agents for Open-World Mobile Manipulation},
  author = {Mi, Boyu and Ma, Mengchen and Yao, Yifei and Gao, Xing and Chen, Junting and Li, Yangzi and Zhu, Zihou and Li, Guohao and Yin, Zhenfei and Wang, Tai and Mu, Yao and Pang, Jiangmiao and Wang, Hanqing},
  booktitle = {European Conference on Computer Vision},
  year = {2026}
}
IROS 20252025

AnyBipe: An Automated End-to-End Framework for Training and Deploying Bipedal Robots Powered by Large Language Models

Yifei Yao, Wentao He, Chenyu Gu, Jiaheng Du, Fuwei Tan, Zhen Zhu, Jun-Guo Lu

Language-guided reward design, reinforcement learning, and sim-to-real evaluation in one end-to-end pipeline for bipedal robots.

BibTeX

Crossref DOI metadata

@inproceedings{Yao_2025,
  title = {AnyBipe: An Automated End-to-End Framework for Training and Deploying Bipedal Robots Powered by Large Language Models},
  url = {http://dx.doi.org/10.1109/IROS60139.2025.11246492},
  DOI = {10.1109/iros60139.2025.11246492},
  booktitle = {2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  publisher = {IEEE},
  author = {Yao, Yifei and He, Wentao and Gu, Chenyu and Du, Jiaheng and Tan, Fuwei and Zhu, Zhen and Lu, Jun-Guo},
  year = {2025},
  month = Oct,
  pages = {14790--14797}
}

Preprints

arXiv2026Co-first author

PICA-WAM: Visual Perturbation Invariance via Cross-Appearance Self-Alignment in World Action Models

Yude Zou*, Yifei Yao*, Guanbo Jiao*, Jianming Ma, Guanjie Zheng, Xing Gao

* Equal contribution (co-first authors).

Cross-appearance self-alignment encourages world action models to produce consistent action representations under task-irrelevant visual changes.

BibTeX

arXiv export

@misc{zou2026picawamvisualperturbationinvariance,
  title = {PICA-WAM: Visual Perturbation Invariance via Cross-Appearance Self-Alignment in World Action Models},
  author = {Yude Zou and Yifei Yao and Guanbo Jiao and Jianming Ma and Guanjie Zheng and Xing Gao},
  year = {2026},
  eprint = {2610.26813},
  archivePrefix = {arXiv},
  primaryClass = {cs.CV},
  url = {https://arxiv.org/abs/2610.26813}
}
SSRN2026

Length-Adaptive PPO-based RLHF for Non-Uniform UAV Control

Fuwei Tan, Zhen Zhu, Yifei Yao, Qing-Hao Zhang, Xiaodong Han, Junguo Lu

Length-adaptive proximal policy optimization with reinforcement learning from human feedback for non-uniform UAV control.

BibTeX

Crossref DOI metadata

@misc{Tan_2026,
  title = {Length-Adaptive PPO-based RLHF for Non-Uniform UAV Control},
  url = {http://dx.doi.org/10.2139/ssrn.6085506},
  DOI = {10.2139/ssrn.6085506},
  publisher = {Elsevier BV},
  author = {Tan, Fuwei and Zhu, Zhen and Yao, Yifei and Zhang, Qing-Hao and Han, Xiaodong and Lu, Junguo},
  year = {2026},
  note = {SSRN preprint}
}
arXiv2025

GBC: Generalized Behavior-Cloning Framework for Whole-Body Humanoid Imitation

Yifei Yao, Chengyuan Luo, Jiaheng Du, Wentao He, Jun-Guo Lu

An open-source pathway from human motion to humanoid action: differentiable retargeting, DAgger-MMPPO, and an Isaac Lab training workflow.

BibTeX

arXiv export

@misc{yao2025gbcgeneralizedbehaviorcloningframework,
  title = {GBC: Generalized Behavior-Cloning Framework for Whole-Body Humanoid Imitation},
  author = {Yifei Yao and Chengyuan Luo and Jiaheng Du and Wentao He and Jun-Guo Lu},
  year = {2025},
  eprint = {2508.09960},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO},
  url = {https://arxiv.org/abs/2508.09960}
}

Technical Reports

arXiv2026Co-first author

Profiling a Frontier Agent with EBench: From Direct Control to Harness with Policy

Yifei Yao*, Yude Zou*, Guanbo Jiao*, Xing Gao, …, Haoyu Guo, Weinan Zhang, Chunhua Shen

* Equal contribution (co-first authors).

Evaluating a frontier agent on EBench under direct robot control and with a harness that can invoke post-trained robot policies.

BibTeX

arXiv export

@misc{yao2026profilingfrontieragentebench,
  title = {Profiling a Frontier Agent with EBench: From Direct Control to Harness with Policy},
  author = {Yifei Yao and Yude Zou and Guanbo Jiao and Xing Gao and Ning Gao and Yukai Wang and Jiantong Chen and Zihou Zhu and Xinyu Li and Ningyu Yan and Shuai Wang and Jianming Ma and Haoyu Guo and Weinan Zhang and Chunhua Shen},
  year = {2026},
  eprint = {2610.35832},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO},
  url = {https://arxiv.org/abs/2610.35832}
}
Technical report2026

InternW0-Δ: A World Action Model Bridging Predictive Dynamics and Actions with 20K+ Hours of Open Data

Xingyu Miao, Zizun Li, …, Yifei Yao, …, Chunhua Shen

A world action model combining predictive dynamics, semantic guidance, and geometric priors, pretrained on more than 20,000 hours of open data.

BibTeX

arXiv export

@misc{miao2026internw0deltaworldactionmodel,
  title = {InternW0-$\Delta$: A World Action Model Bridging Predictive Dynamics and Actions with 20K+ Hours of Open Data},
  author = {Xingyu Miao and Zizun Li and Baole Fang and Kaiwen Song and Tenghui Wang and Hanxue Zhang and Yating Wang and Xudong Li and Yuping He and Xueyuan Wei and Chao Gao and Xijie Yang and Yingxiang Xu and Kerui Ren and Wenqi Guo and Jianjun Zhou and Xinzhe Wang and Weiguang Zhao and Ni Yang and Zetao Cai and Yufei Xue and Hengjie Li and Zeyu He and Yuanzhen Zhou and Rong Fu and Jianyang Zhang and Siwei Cui and Fuxian Huang and Yunsong Zhou and Xing Gao and Yifei Yao and Qiaojun Yu and Kailin Li and Ming Zhou and Mu Huang and Xinyue Li and Wenze Cui and Bingqi Jiang and Xueyue Zhu and Junting Dong and Haoyu Guo and Tao Lu and Mulin Yu and Bowen Zhou and Bin Zhao and Tianfan Xue and Weinan Zhang and Chunhua Shen},
  year = {2026},
  eprint = {2609.31394},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO},
  url = {https://arxiv.org/abs/2609.31394}
}
Technical report2026

InternW0: A Foundational Physical World Model for Efficient Real-World Interactions

Jisong Cai, Yao Mu, …, Yifei Yao, …, Weinan Zhang

An asymmetric video-action architecture couples future visual dynamics with continuous robot control, supporting asynchronous prediction and contact-aware interaction.

BibTeX

arXiv export

@misc{cai2026internw0foundationalphysicalworld,
  title = {InternW0: A Foundational Physical World Model for Efficient Real-World Interactions},
  author = {Jisong Cai and Yao Mu and Ganlin Yang and Zhe Cao and Zhangzheng Tu and Xing Gao and Kailin Li and Xinyu Zhan and Lixin Yang and Yangkun Zhu and Haoxiang Ma and Ming Zhou and Qiaojun Yu and Yufei Xue and Liqun He and Yifei Yao and Yifan Zhu and Long Ling and Bingqi Jiang and Haoyu Guo and Xueyue Zhu and Bowen Zhou and Bin Zhao and Tianfan Xue and Chunhua Shen and Weinan Zhang},
  year = {2026},
  eprint = {2609.27656},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO},
  url = {https://arxiv.org/abs/2609.27656}
}

Ongoing Research

Research project2026 · In progress

Learning Embodied Agents from Experience

How can an open model learn to act from a frontier agent’s experience? I am studying the transfer of Astra’s robot interaction experience to Qwen3.8-27B through staged SFT and reinforcement learning, connecting visual grounding, short manipulation skills, and long-horizon tasks.

Research project2026 · In development

ARIEL: An Agent-Ready Humanoid Behavior Model with In-Context Execution Memory

ARIEL gives an agent a simple interface to whole-body control. Choose a demonstration, load it into memory, then recall the motion or steer its style with velocity and heading commands. The controller keeps executing between agent decisions.

Watch demos

Three motions, replayed from memory

Three independent simulation trials: disco, reaching overhead and a questioning gesture. Loaded demonstrations are replayed without motion references. Actions remain at original speed; the last two trials explicitly omit four seconds of waiting.

Keep a style while changing direction

A separate simulation trial: a preloaded walking style follows changing velocity and heading commands. One continuous 28-second segment at original speed.
Preprint2026 · arXiv release pendingCo-first author

Review the Episodes, Refine the Lessons: Simple Memory Evolution for Embodied Agents

Yifei Yao*, Yude Zou*, Guanbo Jiao, Yukai Wang, Jianming Ma, Xing Gao, Weinan Zhang, Chunhua Shen

* Equal contribution (co-first authors).

An embodied agent learns from execution experience by refining text memory across training episodes, then uses frozen memory at evaluation without updating model weights.

Manuscript2026 · Manuscript in preparation

ABI: Actor-Free Zero-Shot Reinforcement Learning with Behavior-Inductive Representations

Yifei Yao

First author; full author list to be finalized.

Given an existing policy, ABI trains only a reusable guide. The guide improves task execution at deployment while the policy stays frozen, with no parameter updates or task-specific policy post-training.

Watch demo

Same policy, with and without ABI

Manipulation, locomotion and navigation excerpts from the project video. Each pair uses the same starting state, random seed and frozen base policy. Selected simulation examples with edited playback speeds; on-screen scores describe these development episodes.