Pengfei Li

Pengfei Li

Ph.D. Student, Embodied AI & Autonomous Driving

Institute for AI Industry Research (AIR), Tsinghua University · Beijing, China

About

I am a fourth-year Ph.D. student at the Institute for AI Industry Research (AIR), Tsinghua University, advised by Prof. Yilun Chen and Prof. Ya-Qin Zhang. I received my B.S. from the University of Chinese Academy of Sciences (UCAS) in 2022, where I worked with Prof. Qingming Huang.

My research focuses on Embodied AI and Autonomous Driving, with the goal of developing, optimizing, and deploying AI algorithms in real-world industrial scenarios. I am especially interested in humanoid whole-body control, loco-manipulation, and world model.

Selected Publications (* equal contribution)

First-author / co-first-author works are highlighted below. A full list — including collaborations — is at the bottom of this section.

Rhythm teaser
RSS 2026 Rhythm: Learning Interactive Whole-Body Control for Dual Humanoids

Hongjin Chen*, Wei Zhang*, Pengfei Li*, Shihao Ma, Ke Ma, Yujie Jin, Zijun Xu, Xiaohui Wang, Yupeng Zheng, Zining Wang, Jieru Zhao, Yilun Chen, Wenchao Ding.

First unified framework for real-world dual-humanoid whole-body interaction — interaction-aware retargeting + interaction-guided RL — deployed on Unitree G1 (greeting, hugging, dancing).

Paper Project

Human and algorithmic visual attention teaser
npj Artificial Intelligence 2026 Human and Algorithmic Visual Attention in Driving Tasks

Chen Zheng*, Pengfei Li*, Bu Jin*, Shanhe You, Ka I Chan, Ya-Qin Zhang, Guyue Zhou, Jiangtao Gong.

Decomposes human driving attention into spatial / feature-based / mixed phases and shows that injecting human semantic attention closes both the "reasoning" and "grounding" gaps of detection, planning, and VLM models.

Paper

Dual-AEB teaser
ICRA 2025 Dual-AEB: Synergizing Rule-Based and Multimodal Large Language Models for Effective Emergency Braking

Wei Zhang*, Pengfei Li*, Junli Wang, Bingchuan Sun, Qihao Jin, Guangjun Bao, Shibo Rui, Yang Yu, Wenchao Ding, Peng Li, Yilun Chen.

First MLLM-augmented AEB system — combines fast rule-based reaction with rich open-scenario reasoning from a multimodal LLM.

Paper Code

LiON teaser
AAAI 2025 LiON: Learning Point-wise Abstaining Penalty for Point Cloud Anomaly Detection

Shaocong Xu*, Pengfei Li*, Xinyu Liu, Qianpu Sun, Yang Li, Shihui Guo, Zhen Wang, Bo Jiang, Rui Wang, Kehua Sheng, Bo Zhang, Hao Zhao.

Reframes LiDAR outlier detection as selective classification with a learned point-wise abstaining penalty and synthetic-outlier curriculum — SOTA on SemanticKITTI and nuScenes.

Paper Code

LODE teaser
ICRA 2023 LODE: Locally Conditioned Eikonal Implicit Scene Completion from Sparse LiDAR

Pengfei Li, Ruowen Zhao, Yongliang Shi, Hao Zhao, Jirui Yuan, Guyue Zhou, Ya-Qin Zhang.

An eikonal-constrained implicit representation that turns sparse outdoor LiDAR into dense semantic 3D scenes — the first locally conditioned eikonal completion for autonomous driving.

Paper Code

TOIST teaser
NeurIPS 2022 TOIST: Task Oriented Instance Segmentation Transformer with Noun-Pronoun Distillation

Pengfei Li, Beiwen Tian, Yongliang Shi, Xiaoxue Chen, Hao Zhao, Guyue Zhou, Ya-Qin Zhang.

Formulates task-oriented detection as preferred-object segmentation, and introduces noun-to-pronoun distillation so a transformer can act on verbs like "sit on" without explicit object naming.

Paper Code

More publications

→ See full list on Google Scholar

Awards & Honors

  • 20231st Place — ICRA 2023 PUB.R Competition (Preparation and Dish-Up of an English Breakfast with Robots).
  • 2021National Scholarship (Undergraduate, Top 1%).
  • 2020National Scholarship (Undergraduate, Top 1%).

Education

  • 2022 – PresentPh.D. Student, Institute for AI Industry Research (AIR), Tsinghua University.
  • 2018 – 2022B.S., School of Computer Science and Technology, University of Chinese Academy of Sciences.
    • GPA 3.94 / 4.00, Rank 1 / 104.