Scholar
Tianhao Wang
Google Scholar ID: m45LD1kAAAAJ
Assistant Professor, University of California, San Diego
Citations & Impact
Citations
858
H-index
18
i10-index
22
Publications
20
Co-authors
31
list available
Publications
Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
- - Paper: 'Provable Benefit of Sign Descent: A Minimal Model Under Heavy-Tail Class Imbalance', NeurIPS 2025 Workshop on Optimization for Machine Learning (OPT 2025), Oral Presentation
- - Paper: 'On Universality of Non-Separable Approximate Message Passing Algorithms', arXiv:2506.23010, 2025
- - Paper: 'Taming Polysemanticity in LLMs: Provable Feature Recovery via Sparse Autoencoders', arXiv:2506.14002, 2025
- - Paper: 'Structured Preconditioners in Adaptive Optimization: A Unified Analysis', International Conference on Machine Learning (ICML), 2025
- - Paper: 'Can Neural Networks Achieve Optimal Computational-statistical Tradeoff? An Analysis on Single-Index Model', International Conference on Learning Representations (ICLR), 2025, Oral Presentation
- - Paper: 'How well can Transformers emulate in-context Newton’s method?', International Conference on Artificial Intelligence and Statistics (AISTATS), 2025
- - Paper: 'Unveiling Induction Heads: Provable Training Dynamics and Feature Learning in Transformers', Advances in Neural Information Processing Systems (NeurIPS), 2024
- - Paper: 'Implicit regularization of gradient flow on one-layer softmax attention', arXiv:2403.08699, 2024
- - Paper: 'Approximate Message Passing for orthogonally invariant ensembles: Multivariate non-linearities and spectral initialization', Information and Inference: A Journal of the IMA, 2024
- - Paper: 'Universality of Approximate Message Passing algorithms and tensor networks', Annals of Applied Probability, 2024
Research Experience
- - Research Assistant Professor, Toyota Technological Institute at Chicago, 2024-2025, Collaborators: Zhiyuan Li and Nathan Srebro
Education
- - Research Assistant Professor, Toyota Technological Institute at Chicago, 2024-2025, Advisors: Zhiyuan Li and Nathan Srebro
- - Ph.D., Department of Statistics and Data Science, Yale University, Advisor: Zhou Fan
Background
- Currently an Assistant Professor at the Halıcıoğlu Data Science Institute at the University of California, San Diego. Broadly interested in various aspects of machine learning, optimization, and statistics.
