Dr. Shangqian Gao
Assistant Professor
- Office: 251A Love Building
- Telephone:
- Email: sgao [ at fsu dot edu ]
Research
My research interests includes:
- Large Language Models: Model compression for Large Language Models (LLMs), including but not limited to structural pruning, Singular-value decomposition (SVD), layer-wise sparsity allocation, sparse Mixture of Experts (MoE), etc. Memorization in LLMs. Self-improvement and alignment for LLMs.
- Efficient Vision: Differentiable Neural Architecture Search (NAS), and efficient image and image-to-text generation models, such as diffusion models and GANs.
- Policy Gradient Methods for Reinforcement Learning: Variance-reduced policy gradient methods based on momentum techniques and mirror descent.