An Interactive AI Physicist Demonstration for Open Science
- Room: 80/1-001 - Globe of Science and Innovation - 1st Floor
- Speaker:
- Matthias Le Dall, , Matthias obtained his PhD in theoretical physics in 2017 from the University of Victoria, where he studied the theories of particle physics that explain the matter content of the universe, including dark matter. He then turned his attention to quantum computing, where he studied the sources of noise in the superconductors that D-Wave uses to manufacture their quantum bits. He worked at Scotiabank for 6 years developing machine learning and quantum computing algorithms to forecast financial markets and optimize the bank's trading strategies. Today at FirstPrinciples, Matthias enjoys working at the intersection between his two passions: Physics and AI. Here he works to develop an AI framework that will assist physicists do their research, including testing existing theories and making new hypotheses. , FirstPrinciples, https://www.firstprinciples.org/, Canada
- Nawar Ismail, , Nawar completed his undergraduate and master’s degree in physics at the University of Guelph. He studied nuclear physics, focusing on numerical simulations and programming. He was further driven into optimization and automation as a need to manage the intense computational resources required to carry out that research. Since then, he has studied a whole host of interesting problems from game theory, simulations, optimization, automation, reinforcement learning, and data science. , FirstPrinciples, https://www.firstprinciples.org/, Canada
In this interactive demo, participants will have the opportunity to directly interact with our AI system, experiencing firsthand how it navigates scientific questions, autonomously searches for relevant literature, and performs symbolic mathematical reasoning in real-time. A particular highlight will include our system's capability for symbolic regression demonstrated via an interactive scenario, such as deriving equations governing simple physical systems from raw data.
Fundamental physics research today is constrained by unprecedented complexity and information overload. FirstPrinciples is developing an advanced AI system designed to accelerate discovery and enhance transparency in fundamental physics research. Leveraging large-language models, symbolic reasoning engines, and reinforcement learning, our system will ingest open-access scientific datasets and literature to derive equations, generate hypotheses, and synthesize existing research, significantly improving researchers’ productivity and insights.
In this interactive demo, participants will have the opportunity to directly interact with our AI system, experiencing firsthand how it navigates scientific questions, autonomously searches for relevant literature, and performs symbolic mathematical reasoning in real-time. A particular highlight will include our system's capability for symbolic regression demonstrated via an interactive scenario, such as deriving equations governing simple physical systems from raw data.
Participants will observe the system's transparent reasoning process, which clearly displays its internal steps, references to open-access sources, and uncertainty indicators, among other essential features to ensure trust and reproducibility. This demonstration aims to foster meaningful interactions, gathering valuable community feedback to guide future improvements, collaborations, and expansions.
We anticipate this interactive session will facilitate networking, encourage collaboration, highlight the transformative potential of transparent AI assistance within the open research community, and demonstrate practical alignment with OSFair’s emphasis on openness, reproducibility, and innovative research infrastructures supporting sustainable scientific collaboration.