Distilling Neuro-Symbolic Programs into 3D Multi-modal LLMs
Wentao Mo, Yang Liu
Open to research collaborations
Peking University / Wangxuan Institute of Computer Technology
Wentao Mo
PhD student at Peking University working on multimodal learning, computer vision, and embodied AI.
I joined the Wangxuan Institute of Computer Technology (WICT), PKU in September 2022 for my PhD, supervised by Prof. Yang Liu. My work explores how vision, language, and 3D scene understanding can support stronger reasoning in embodied environments.
Publications
Selected work
Advancing 3D Scene Understanding with MV-ScanQA Multi-View Reasoning Evaluation and TripAlign Pre-training Dataset
Wentao Mo, Qingchao Chen, Yuxin Peng, Siyuan Huang, Yang Liu
Bridging the Gap between 2D and 3D Visual Question Answering: A Fusion Approach for 3D VQA
Wentao Mo, Yang Liu
3D Vision and Language Pretraining with Large-Scale Synthetic Data
Dejie Yang, Zhu Xu, Wentao Mo, Qingchao Chen, Siyuan Huang, Yang Liu
Research
3D Vision-Language Learning
Fusing 2D visual priors with 3D geometry for scene understanding and practical question answering.
3D Pre-training & Evaluation
Building multi-view evaluations and pre-training data that ground models in 3D environments.
Neuro-Symbolic 3D Reasoning
Distilling verified neuro-symbolic programs into natural-language reasoning, so 3D MLLMs can show how they think.
From 3D perception to grounded reasoning
Contact