Publications
* equal contribution. Also see my CV.
ChairNav: Cross-Embodiment Pre-training and Personalization for Long-Horizon Wheelchair Navigation
Under submission, 2026
Adapts a visual navigation foundation model to a Permobil M3 power wheelchair through cross-embodiment pre-training, wheelchair mid-training, and human-in-the-loop post-training from user takeovers; 84.9% autonomous driving time across 7.8 km of outdoor navigation. *Equal contribution.
Junhua Huang*, Zhizheng Liu*, Honglin He, Xiang Zhang, Bolei Zhou. (2026). "ChairNav: Cross-Embodiment Pre-training and Personalization for Long-Horizon Wheelchair Navigation." arXiv:2610.24012 (under submission).
Cue the Flow: Steering Flow-Matching Policies for Open-World Delivery Manipulation
Conference on Robot Learning (CoRL 2026), 2026
Steering flow-matching manipulation policies with VLM-generated cues for open-world delivery.
Haoxuan Wang, Griffin Galimi, Junhua Huang, Selina Song, Wayne Wu, Yan Yan, Bolei Zhou. (2026). "Cue the Flow: Steering Flow-Matching Policies for Open-World Delivery Manipulation." Conference on Robot Learning (CoRL 2026).
ToolSmith: An Automatic Tool-Generation Framework for Multimodal Search Agents
Findings of EMNLP 2026, 2026
A self-expanding multimodal search agent that automatically generates and verifies task-specific tools.
Rui Sun, Junhua Huang, Luca Zancato, Md Mofijul Islam, Kai-Wei Chang. (2026). "ToolSmith: An Automatic Tool-Generation Framework for Multimodal Search Agents." Findings of EMNLP 2026.
[Project]
Seizure-Semiology-Suite (S3): A Clinically Multimodal Dataset, Benchmark, and Models for Seizure Semiology Understanding
ICML 2026 (Spotlight), 2026
A clinically multimodal dataset, benchmark, and models for seizure semiology understanding.
Lina Zhang, Tonmoy Monsoor, et al. (incl. Junhua Huang), Rajarshi Mazumder, Vwani Roychowdhury. (2026). "Seizure-Semiology-Suite (S3): A Clinically Multimodal Dataset, Benchmark, and Models for Seizure Semiology Understanding." ICML 2026 (Spotlight).
[Paper]
GestureLSM: Latent Shortcut based Co-Speech Gesture Generation with Spatial-Temporal Modeling
ICCV 2025, 2025
Latent Shortcut based Co-Speech Gesture Generation with Spatial-Temporal Modeling.
Pinxin Liu, Luchuan Song, Junhua Huang, Haiyang Liu, Junfan Zhu, Chenliang Xu. (2025). "GestureLSM: Latent Shortcut based Co-Speech Gesture Generation with Spatial-Temporal Modeling." ICCV 2025.
[Paper]
Semantic Visually-Guided Acoustic Highlighting with Large Vision-Language Models
arXiv preprint, 2026
Using large VLMs to guide acoustic highlighting from visual context.
Junhua Huang, Chao Huang, Chenliang Xu. (2026). "Semantic Visually-Guided Acoustic Highlighting with Large Vision-Language Models." arXiv preprint.
[Paper]
GestureFAR: Streaming Co-Speech Gesture Generation with Flow Autoregression
arXiv preprint, 2026
Streaming co-speech gesture generation with flow autoregression.
Pinxin Liu, Haiyang Liu, Jiahao Luo, Junhua Huang, Chunhao Zou, Luchuan Song. (2026). "GestureFAR: Streaming Co-Speech Gesture Generation with Flow Autoregression." arXiv preprint.
[Paper]
Video-LMM Post-Training: A Deep Dive into Video Reasoning with Large Multimodal Models
arXiv preprint, 2025
A survey of post-training methods for video reasoning with large multimodal models.
Yolo Y. Tang, Jing Bi, Pinxin Liu, et al. (incl. Junhua Huang), Jiebo Luo, Chenliang Xu. (2025). "Video-LMM Post-Training: A Deep Dive into Video Reasoning with Large Multimodal Models." arXiv preprint.
[Paper]
Using Generative AI to Calculate Party Positions: A Comparison of Human Experts and Large Language Models
Journal of Political Institutions and Political Economy, 2025
This paper compares human experts and LLMs in calculating party positions.
Cantay Caliskan, Junhua Huang, Yiyang Huang, Ruoxuan Lin, Wanting Shan. (2025). "Using Generative AI to Calculate Party Positions: A Comparison of Human Experts and Large Language Models." Journal of Political Institutions and Political Economy.