AI is empowering research discoveries in various scientific fields, integrating into the phases of hypothesis formation, experimental design, data collection, and analysis of scientific discovery. The paper elaborates on the applications of technologies such as self-supervised learning, geometric deep learning, and generative artificial intelligence in scientific research, while also highlighting the core issues that remain in interdisciplinary AI development. Three Chinese co-authors collaborated with Yoshua Bengio and his team, and the paper has been published as a review article in Nature.
