Fanghui Liu

Venues

Frequent coauthors

Papers

1. How Gradient descent balances features: A dynamical analysis for two-layer neural networks
   ICLR 2025 ·

   Zhenyu Zhu

   DBLP profile ↗ORCID search ↗
2. LoRA-One: One-Step Full Gradient Could Suffice for Fine-Tuning Large Language Models, Provably and Efficiently
   ICML 2025 ·

   Yuanhe Zhang

   DBLP profile ↗ORCID search ↗
3. The φ Curve: The Shape of Generalization through the Lens of Norm-based Capacity Control
   NeurIPS 2025 ·

   Yichen Wang

   DBLP profile ↗ORCID search ↗
4. Efficient local linearity regularization to overcome catastrophic overfitting
   ICLR 2024 ·

   Elías Abad-Rocamora

   DBLP profile ↗ORCID search ↗
5. Generalization of Scaled Deep ResNets in the Mean-Field Regime
   ICLR 2024 ·

   Yihang Chen

   DBLP profile ↗ORCID search ↗
6. High-Dimensional Kernel Methods under Covariate Shift: Data-Dependent Implicit Regularization
   ICML 2024 ·

   Yihang Chen

   DBLP profile ↗ORCID search ↗
7. Learning Scalable Model Soup on a Single GPU: An Efficient Subspace Training Strategy
   ECCV 2024 ·

   Tao Li

   DBLP profile ↗ORCID search ↗
8. Revisiting Character-level Adversarial Attacks for Language Models
   ICML 2024 ·

   Elías Abad-Rocamora

   DBLP profile ↗ORCID search ↗
9. Robust NAS under adversarial training: benchmark, theory, and beyond
   ICLR 2024 ·

   Yongtao Wu

   DBLP profile ↗ORCID search ↗
10. The Role of Over-Parameterization in Machine Learning - the Good, the Bad, the Ugly
    AAAI 2024 · Fanghui Liu