November 22, 2024

The Next Einstein: New AI Can Develop New Theories of Physics

Their AI is able to acknowledge patterns in complicated data sets and to develop them in a physical theory.In the following interview, Prof. Moritz Helias from Forschungszentrum Jülichs Institute for Advanced Simulation (IAS-6) explains what the “Physics of AI” is all about and to what extent it varies from traditional approaches.How do physicists come up with a brand-new theory?You generally start with observations of the system before trying to propose how the different system components interact with each other in order to discuss the observed habits. In the future, much larger systems need to likewise be possible through further optimization.How does this approach differ from other AIs such as ChatGPT?Many AIs objective to discover a theory of the information utilized to train the AI. The theories that the AIs learn typically can not be analyzed. It hence belongs to the field of explainable AI, specifically the “physics of AI”, as we use the language of physics to discuss what the AI has actually learned. We can use the language of interactions to develop a bridge between the complex inner functions of AI and theories that people can understand.Reference: “Learning Interacting Theories from Data” by Claudia Merger, Alexandre René, Kirsten Fischer, Peter Bouss, Sandra Nestler, David Dahmen, Carsten Honerkamp and Moritz Helias, 20 November 2023, Physical Review X.DOI: 10.1103/ PhysRevX.13.041033.

Their AI is able to recognize patterns in complex data sets and to develop them in a physical theory.In the following interview, Prof. Moritz Helias from Forschungszentrum Jülichs Institute for Advanced Simulation (IAS-6) discusses what the “Physics of AI” is all about and to what extent it differs from traditional approaches.How do physicists come up with a new theory?You normally begin with observations of the system before trying to propose how the different system elements connect with each other in order to discuss the observed behavior. In the future, much larger systems must also be possible through additional optimization.How does this approach differ from other AIs such as ChatGPT?Many AIs objective to learn a theory of the information used to train the AI. It therefore belongs to the field of explainable AI, particularly the “physics of AI”, as we utilize the language of physics to describe what the AI has discovered.