NUS ECE is proud to announce that Ms. Ma Xinyin, a PhD student in our programme, has been awarded the Google PhD Fellowship, one of the most prestigious accolades for doctoral students in computer science and engineering worldwide.
This distinguished fellowship honors exceptional PhD candidates who demonstrate outstanding potential in shaping the future of technology. Xinyin’s research focuses on advancing the efficiency of deep learning models, including large language models (LLMs) and diffusion models, a critical area of innovation in the field of artificial intelligence.
Under the guidance of Assistant Professor Wang Xinchao and in collaboration with her lab mates, Xinyin has developed several widely recognized compression algorithms to accelerate generative models. Among her notable contributions is LLM-Pruner, a pioneering structural pruning framework for large language models. As one of the world’s first lightweight LLMs, LLM-Pruner has had a significant impact on the AI community and beyond, earning over 380 citations within a year of publication. Another standout innovation from Xinyin’s research is DeepCache, which introduces training-free compression methods for diffusion models. This cutting-edge work has been integrated into TensorRT and Diffusers, the world’s leading frameworks for high-performance deep learning and image generation.
Xinyin’s trailblazing contributions to AI research have redefined the efficiency of complex AI models while maintaining their high performance. Her work paves the way for deploying large-scale AI models on edge devices, making advanced AI technologies more accessible and versatile.
Congratulations to Ms. Ma Xinyin on this outstanding achievement!
