Scholar
Yixiao HUANG
Google Scholar ID: iTEcewwAAAAJ
PhD student at EECS, University of California, Berkeley
Citations & Impact
Citations
85
H-index
4
i10-index
2
Publications
10
Co-authors
15
list available
Publications
Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
- Publications:
- - OVERT: A Benchmark for Over-Refusal Evaluation on Text-to-Image Models (NeurIPS 2025, D&B Track)
- - Generalization or Hallucination? Understanding Out-of-Context Reasoning in Transformers (NeurIPS 2025)
- - Understanding and Improving Fast Adversarial Training against ℓ0 Bounded Perturbations (NeurIPS 2025)
- - Reinforcement Learning for Flow-Matching Policies (arXiv preprint, 2025)
- - On the Power of Convolution Augmented Transformer (AAAI 2025)
- - From Self-Attention to Markov Models: Unveiling the Dynamics of Generative Transformers (ICML 2024)
- - Mechanics of Next Token Prediction with Self-Attention (AISTATS 2024)
Research Experience
- Worked with Prof. Samet Oymak at the University of Michigan on the theoretical foundations of self-attention and hybrid models.
Education
- Bachelor's Degree: Graduated from the Department of Electrical Engineering, City University of Hong Kong in June 2023, supervised by Prof. Rosa CHAN; Currently a second-year PhD student at UC Berkeley EECS, supervised by Prof. Somayeh Sojoudi, and also works closely with Prof. Song Mei.
Background
- Research Interests: Establishing the theoretical foundations for generative models to enhance their efficiency and reliability. Field: Information Engineering.
Miscellany
- Looking for a research internship in Summer 2026.
