California Institute of Technology and the University of Chicago co-hosted the 2nd AI+Science Conference on November 10 and 11, 2025. The meeting was held at the Institute. Huntington. And online, we spotlighted the potential of artificial intelligence (AI) and machine learning to accelerate scientific discovery across the physical and biological sciences. The conference, funded by the Margot and Tom Pritzker Foundation, also included the announcement of the inaugural recipients of the Margot and Tom Pritzker Prize for Excellence in AI in Scientific Research.
“Over the years, Margot and I have come to believe that science is where our philanthropy can have the most lasting impact,” said Tom Pritzker, co-founder of the Margot and Tom Pritzker Foundation. “We believe science is one of the most powerful tools for progress. Early and wise risks can create breakthroughs whose benefits resonate across generations. We believe in enabling great people and the ecosystems around them. At the Pritzker Science Prize, we want to celebrate the curiosity, rigor, and tenacity that moves science forward, knowing that every insight, successful or not, expands the possibilities for all of us.”
Conference speakers discussed research in physics, biology, health, neuroscience, climate, and robotics that is being taken on new avenues by advances in AI algorithms. ”
Videos of all conference presentations are now available on the conference website.
Anima Anandkumar, Caltech’s Bren Professor of Computing and Mathematical Sciences and conference co-sponsor, gave opening remarks thanking the Pritzers for their vision and generosity in bringing Caltech and the University of Chicago together for this unique event. She then summarized the rich history of AI at Caltech. Its history goes back to an early course on computational physics taught by the late Richard Feynman. Carver Mead (BS ’56, PhD ’60), Gordon and Betty Moore Professor Emeritus of Engineering and Applied Science. John Hopfield, Roscoe G. Dickinson Professor Emeritus of Chemistry and Biology at the California Institute of Technology and co-recipient of the 2024 Nobel Prize in Physics. This includes the creation of the AI+Science Initiative in 2018.
Anandkumar then shared his vision for building AI that pushes the frontiers of scientific discovery, noting that a central bottleneck in the research process is the need to conduct physical experiments. Anandkumar shared her research on “neural operators.” This is a type of machine learning architecture that can incorporate a type of physical understanding into AI, allowing models to simulate, design, and control physical experiments. These neural operators are already having great effects in areas such as weather forecasting, nuclear fusion, and medical device design.
On the first day of the conference, Margot Pritzker announced the winners of the Margot Pritzker Prize and Tom Pritzker Prize for Excellence in AI Scientific Research. Kyle Cranmer, professor of physics at the University of Wisconsin-Madison, and David R. Anderson, director of the University’s Data Science Institute. and Deborah Marks, professor of systems biology at Harvard Medical School. In his talk, Cranmer described the role machine learning has played at CERN’s Large Hadron Collider (LHC) in Geneva, Switzerland, including the discovery of the Higgs boson, and Marks discussed how AI is revolutionizing human genetics.
“Selecting Kyle and Debbie was not an easy task,” said Rebecca Willett, co-chair of the AI+Science Initiative. Willett is director of the AI Department at the University of Chicago Data Science Institute. “But I think what sets their work apart in this crowded and exciting field is that they’re not only using AI to analyze scientific data, but they’re actually identifying areas where current AI methodologies are inadequate for scientific goals, expanding the scope of AI tools, developing new tools, and really advancing both the scientific and AI fields.”
The conference focused on physics and how AI is being used to advance weather and climate modeling and featured the following talks:
- Pedram Hassanzadeh from the University of Chicago, Chris Bretherton from Seattle’s Allen Institute for Artificial Intelligence (Ai2), Laure Zanna from New York University, and Ashesh Chattopadhyay from the University of California, Santa Cruz, discuss recent big advances in AI weather modeling. It started with FourCastNet, the first AI-based weather model launched almost four years ago. AI weather models are currently running in major weather agencies and are deployed during important weather events, for example to help Indian farmers prepare for India’s monsoon season. The researchers noted that long-term climate modeling requires careful modeling of fine-scale effects. Atmospheric model created by Ai2 ACE2 is a state-of-the-art AI-based climate emulator. By combining this model with an AI-based ocean model, Samudra, created by a multi-institutional consortium called M2LiNES, has enabled the development of coupled ocean-atmosphere models for climate.
- Jane Bay, assistant professor of aerospace and Susan Wu Scholar at Caltech, highlighted how AI has revolutionized fluid mechanics in difficult situations posed by turbulence. One such method is reinforcement learning to reduce drag in such situations.
- Fermilab’s Jennifer Ngadiuba highlighted the need for AI and introduced the first automatic encoder-based anomaly detection at the LHC.
- Soon-Jo Chung, Bren Professor of Control and Dynamic Systems at Caltech and senior research scientist at JPL, which Caltech manages for NASA, discussed the importance of protocols to ensure safety in applications such as drone flight in turbulence.
- Yuke Zhu from the University of Texas at Austin and NVIDIA described the use of physics-based simulation to train generalist robots.
The conference also highlighted the role AI has played in chemistry and biology, beyond well-known applications such as modeling protein folding (work recognized with the 2024 Nobel Prize in Chemistry).
In a conference talk, Francis Arnold, the Linus Pauling Professor of Chemical Engineering, Bioengineering, and Biochemistry, director of the Donna and Benjamin M. Rosen Center for Bioengineering and co-recipient of the 2018 Nobel Prize in Chemistry at Caltech, explained how evolution and AI can synergize in overcoming difficult molecular optimization environments to create new functional enzymes that are active, versatile, and capable of catalyzing a variety of reactions. Compatible with a wide range of industrial processes.
In other talks, Arvind Ramanathan of Argonne National Laboratory discussed the wide range of ways AI is helping biology, including protein and genomic language models that can design new enzymes and predict variants of concern during a pandemic, robots that automate physical experiments, and text models and agents for scientific reasoning, while Richard Andersen, James G. Boswell Professor of Neuroscience and director of the T&C Chen Brain Machine Interface Center at Caltech, discussed brain-machine interfaces and AI Learn about the latest developments in neuroscience, including how is improving real-time decoding by reducing frame rate requirements.
Additionally, Srinivas Turaga of the Howard Hughes Medical Institute’s Janelia Research Campus described advances in understanding the fruit fly nervous system using neural connections and behavioral data, and Wei Gao, professor of medical engineering at Caltech and a researcher at the Heritage Medical Research Institute, discussed the role of AI in medical diagnostics and biosensors, including the use of sweat sensors to monitor various health conditions. Gianmarco Pinton from the University of North Carolina has demonstrated that physics-based machine learning based on ultrasound data is a paradigm shift towards validated and accurate lung ventilation mapping.
At the end of the first day, conference attendees were invited to Huntington for additional activities, including a short dinner talk by three Caltech professors. Alireza Marandi, professor of electrical engineering and applied physics, discussed nonlinear photonics design. Marco Bernardi, professor of applied physics, physics and materials science, spoke about modeling quantum systems.
Conference attendees included students and faculty from the California Institute of Technology and the University of Chicago. Representatives from charities such as Schmidt Sciences, the Kavli Foundation, and Google.org. federal agencies such as the Defense Advanced Research Projects Agency; This includes researchers and engineers from private companies such as Meta, Alphabet, Altos Labs, and NVIDIA.
“We were excited to bring together people from different scientific disciplines who see AI as a unifying force that will foster further collaboration in the future,” says Anandkumar.
