Abstract: Language models have emerged as potent tools with demonstrated efficacy in diverse fields such as natural language processing and multi-modal applications. Recently, the application of language models in the realm of protein structure prediction has shown promising results. In this presentation, we delve into the development and utilization of language models and their effectiveness in predicting molecule structures. By integrating these large-scale sequence data with language models, we aim to unlock new possibilities for advancing the precision of structural predictions for both proteins and RNA molecules.
Speaker's Bio: Siqi Sun is a Young Researcher at Fudan University's Research Institute of Intelligent and Complex Systems and the AI For Science Group at the Shanghai Artificial Intelligence Laboratory. He completed his Ph.D. at the TTIC under Professor Jinbo Xu and earned his Bachelor's from Fudan University. Between 2018 and 2022, he contributed to research on language model and its applications at Microsoft Research. Siqi's work in deep learning spans life sciences and natural language processing, and has published more than 20 papers on top conferences and journals. One of his co-developed algorithms notably won the "Breakthrough/Innovation" award from PLOS Computational Biology in 2018. His work has also caught the attention of prominent media outlets including Science, The New York Times, Adweek, The Register, and Synced.