Portrait of Xiao Wang
Multimodal pre-training at Meta

Xiao Wang

AI Research Scientist @ Meta

CS Ph.D. @ Purdue University · 

Current focus

Pre-training of Multimodal LLMs Multimodal Embedding Vision Tokenizer

About me

Xiao Wang is an AI Research Scientist at Meta, focusing on pre-training of multimodal LLMs, multimodal embedding, and vision tokenizers. Before joining Meta, he was a Member of Technical Staff at xAI, working on vision and multimodal pre-training (Grok 4.5-4.6 vision tokenizer).

He received his Ph.D. in Computer Science from Purdue University, where he focused on representation learning, and his bachelor's degree in Computer Science from Xi'an Jiaotong University. His broader research interests include multimodal pre-training, representation learning, and generative modeling.

Google Scholar citations

Selected publications

* equal contribution  ·  † corresponding author

2026
AutoResearchClaw framework overview
arXiv 2026

AutoResearchClaw: Self-Reinforcing Autonomous Research with Human-AI Collaboration

Jiaqi Liu, Shi Qiu, Mairui Li, Bingzhou Li, Haonian Ji, Siwei Han, Xinyu Ye, Peng Xia, Zihan Dong, Meng Chen, Congyu Zhang, Letian Zhang, Guiming Chen, Haoqin Tu, Xinyu Yang, Lu Feng, Xujiang Zhao, Haifeng Chen, Jiawei Zhou, Xiao Wang, Weitong Zhang, Hongtu Zhu, Yun Li, Jieru Mei, Hongliang Fei, Jiaheng Zhang, Linjie Li, Linjun Zhang, Yuyin Zhou, Sheng Wang, Caiming Xiong, James Zou, Zeyu Zheng, Cihang Xie, Mingyu Ding, Huaxiu Yao

LLM & Agents
2025
2024
2023
2022
2021
2020
2019
2017
Deep learning car-following model
IEEE T-ITS 2017 IEEE Transactions on Intelligent Transportation Systems

Capturing Car-Following Behaviors by Deep Learning

Xiao Wang, Rui Jiang, Li Li, Yilun Lin, Xinhu Zheng, Fei-Yue Wang

Nominated for the George N. Saridis Best Transactions Paper Award

Intelligent Transportation

No publications match this topic.

Open source

Live GitHub star counts

Unified autoregressive large multimodal model with discrete visual tokens for understanding, text-to-image generation, and editing.

GitHub stars

Implicit neural representation hyper-network that produces compact, unified vision embeddings for both recognition and generation (ECCV 2026).

GitHub stars

Self-reinforcing multi-agent system for autonomous research with human-AI collaboration.

GitHub stars

Sample-specific Language Model Optimization at Test-time.

GitHub stars

Rectified-flow-based time series generation.

GitHub stars

MAE+ pretraining for vision transformers with adversarial positional embeddings.

GitHub stars

Shows experimentally that asymmetry matters for Siamese representation learning, with clear gains.

GitHub stars

Cooperative-adversarial contrastive learning where both positive and negative samples are directly learnable.

GitHub stars

General contrastive learning framework that brings in information from stronger augmentations.

GitHub stars

Self-supervised learning through adversarial training of negative examples.

GitHub stars

Self-supervised event boundary detection.

GitHub stars

Semi-supervised learning via self-trained ensemble auto-encoding transformations.

GitHub stars GitHub stars