Guangming Tan

Mastering Sparse CUDA Generation through Pretrained Models and Deep Reinforcement Learning

MatRIS: Toward Reliable and Efficient Pretrained Machine Learning Interatomic Potentials

Learning Fast and Accurate Machine Learning Force Fields via Joint Atomic Energy and Energy Hessian Distillation

ElasticMM: Efficient Multimodal LLMs Serving with Elastic Multimodal Parallelism

Exploring Landscapes for Better Minima along Valleys

ELoRA: Low-Rank Adaptation for Equivariant GNNs