Overview
The project you will work on lies on the boundary between AI and theoretical physics. Physics has been a source of inspiration for innovation since the early days of machine learning. In particular, building on recent developments on VAE and diffusion models, we focus on the role of physics in generative models. In generative AI, images, text or other data such as molecules are generated by learning to invert the coarse graining steps in a diffusion process. This is to be compared with the emergence of new physics at different scales that can be well modeled using the Renormalization Group. We aim at exploiting this analogy further to gain a deeper understanding of generative models on the one hand, and to improve the computational aspects of RG flows on the other. The project will be supervised by Dr Miranda Cheng and partially co-supervised by Prof. Max Welling.
What are you going to do?
You will work together with a strong, welcoming and collaborative teams on both theoretical… physics and machine learning. You will write scientific publications in peer-reviewed scientific journals, present results in leading international conferences, and help supervise Master students.
Tasks and responsibilities:
• Conducting independent research in physics and machine learning, resulting in academic publications in peer-reviewed international journals;
• Help to guide Master students in the group
Company:
University of Amsterdam (UvA)
Qualifications:
Language requirements:
Specific requirements:
Educational level:
Level of experience (years):
Senior (5+ years of experience)
