Accelerating linear algebra with graph neural networks Pioneering the use of graph neural networks as a computational primitive for accelerating numerical algorithms. Self-driving labs for accelerated materials science Building autonomous laboratories that combine robotics and AI into closed-loop systems for rapid experimentation and exploration. Diffusion and generative models Developing and applying diffusion-based generative models for image restoration, conditional inference, and scientific applications. Computational medicine Machine learning and optimization for radiotherapy, medical imaging, and clinical diagnostics.
