Abstract: Over the past five or so years physics researchers have begun using machine learning to significantly aid with various physics research tasks. These include detector design, data analysis, and theoretical research. In this talk, we will mostly focus on ML for data analysis and theory development. We will look at how AI is being used across the various areas of physics, concentrating predominantly on nuclear and particle physics. We look at how the recent step change in capabilities brought about by large language models (LLMs) like GPT4 and Gemini, along with various open source models, have changed the intrinsic nature of physics research forever. We also look at how diffusion and multimodal models are accelerating the pace of scientific discovery and what the future might look like for scientific research as a whole. We will also examine the synergy being developed between the physics and ML communities involving many fruitful collaborations. Come armed with all your questions about, and own experience with, using AI in physics.
Speaker Bio: With a background in particle physics, Peter Morgan has been an AI consultant for over ten years. He is founder of Deep Learning Partnership, an AI consulting company based in London, where he currently advises many companies on how they can gain advantage by leveraging machine learning across their value chain. This includes developing new products and becoming more efficient and competitive across all areas of their organization. Scientific companies he has worked with include those developing products in genomics, solar cells and battery physics, https://www.linkedin.com/in/peter-morgan-8b7ba2/.