I'm a 4th year undergraduate at Columbia University. I'm interested in the potential in building a robust, genuine intelligence in AI, with a focus on multimodal systems that can generalize, adapt, and support personalized users and needs.

Current Research Interests
[0001]
Self-supervised/supervised/reinforcement learning
[0002]
Trust & robustness in human-centered AI
[0003]
Reasoning with different modalities

Publications

2025
ULT

Utility-Learning Tension in Self-Modifying Agents

Charles Wang, Keir Dorchen, Peter Jin
Preprint
arxiv code
2025
MathBode

MathBode: Measuring the Stability of LLM Reasoning using Frequency Response

Charles Wang
NeurIPS 2025 Math-AI Workshop
arxiv code dataset
2026
MI9

MI9: An Integrated Runtime Governance Framework for Agentic AI

Charles Wang, Trisha Singhal, Ameya Kelkar, Jason Tuo
AAAI 2026 TrustAgent Workshop
arxiv code
2025
Zebra-CoT

Zebra-CoT: A Dataset for Interleaved Vision Language Reasoning

Ang Li*, Charles Wang*, Kaiyu Yue*, Zikui Cai*, Ollie Liu*, Deqing Fu*, Peng Guo*, Wang Bill Zhu*, Vatsal Sharan, Robin Jia, Willie Neiswanger, Furong Huang, Tom Goldstein, and Micah Goldblum
Preprint
arxiv code dataset