I am a Ph.D. student in Computer Science at Iowa State University, advised by Prof. Yang Li. I received my M.S. and B.Eng. in Computer Science from Zhejiang University, where I worked with Prof. Zhou Zhao.
My research centers on efficient multimodal models, with broader interests in computer vision and multimodal learning. My long-term goal is to build AI systems that perceive and understand the physical world reliably and efficiently.
-
Chat-Scene++: Exploiting Context-Rich Object Identification for 3D LLM
IEEE TPAMI 2026 ··
-
RoboGround: Robotic Manipulation with Grounded Vision-Language Priors
CVPR 2025 ··
-
Robin3D: Improving 3D Large Language Model via Robust Instruction Tuning
ICCV 2025 ··
-
Chat-3D: Data-Efficiently Learn Large Language Model for Universal 3D Scene Perception
NAACL 2025 ··
-
SpatialCLIP: Learning 3D-aware Image Representations from Spatially Discriminative Language
CVPR 2025 ··
-
GENMANIP: LLM-driven Simulation for Generalizable Instruction-Following Manipulation
CVPR 2025 ··
-
Chat-Scene: Bridging 3D Scene and Large Language Models with Object Identifiers
NeurIPS 2024 ··
-
Grounded 3D-LLM with Referent Tokens
arXiv preprint 2024 ··
-
MMScan: A Multi-Modal 3D Scene Dataset with Hierarchical Grounded Language Annotations
NeurIPS Datasets and Benchmarks 2024 ··
-
FreeBind: Free Lunch in Unified Multimodal Space via Knowledge Fusion
ICML 2024 ··
-
Extending Multi-modal Contrastive Representations
NeurIPS 2024 ··
-
Connecting Multi-modal Contrastive Representations
NeurIPS 2023 ··
-
Distilling Coarse-to-Fine Semantic Matching Knowledge for Weakly Supervised 3D Visual Grounding
ICCV 2023 ··
-
Towards Effective Multi-modal Interchanges in Zero-resource Sounding Object Localization
NeurIPS 2022 ··