The US National Science Foundation (NSF) today announced an investment of more than $ 100 million to establish five institutes of artificial intelligence (AI), each of which will receive approximately $ 20 million over five years. One of these, the NSF AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI), will be led by MIT’s Laboratory for Nuclear Science (LNS) and will be the intellectual home of more than 25 physics and AI senior researchers at MIT and Harvard, Northeast -Easten. , and Tufts universities.
By fusing research into physics and AI, the IAIFI seeks to address some of the most challenging problems in physics, including precision calculations of the structure of matter, detection of gravitational waves from the joining of black holes, and the extraction of new physical laws from loud data,
“IAIFI’s goal is to develop the next generation of AI technologies, based on the transformative idea that artificial intelligence can directly incorporate physics intelligence,” said Jesse Thaler, a physics professor at MIT, LNS researcher, and IAIFI -managing director. “By merging the ‘deep learning’ revolution with the time-tested strategies of ‘deep thinking’ in physics, we strive for a deeper understanding of our universe and of the principles underlying intelligence.”
IAIFI researchers say their approach enables them to make groundbreaking discoveries in physics, and to promote AI more generally, by developing new AI approaches that incorporate first principles from fundamental physics.
“Calling on the simple principle of translation symmetry – which in nature gives rise to conservation of momentum – has led to dramatic improvements in image recognition,” said Mike Williams, a professor of physics at MIT, LNS researcher, and IAIFI deputy director. “We believe that incorporating more complex physics principles will revolutionize how AI is used to study fundamental interactions while advancing the foundations of AI.”
In addition, a core element of the IAIFI mission is to transfer its technologies to the wider AI community.
“NSF recognizes the critical role of AI, and invests in collaborative research and education hubs, such as the NSF IAIFI anchored at MIT, which bring together academia, industry and government to explore deep discoveries and develop new opportunities,” says NSF Director Sethuraman Panchanathan. “Just as previous NSF investments made the breakthroughs that led to today’s AI revolution, the awards announced today will foster discovery and innovation that will sustain U.S. leadership and competitiveness in AI for decades to come. . “
Research in AI and fundamental interactions
Fundamental interactions are described by two pillars of modern physics: at short distances by the Standard Model of Particle Physics, and at long distances by the Lambda Cold Dark Matter model of Big Bang cosmology. Both models are based on physical initial principles such as causality and space-time symmetries. An abundance of experimental evidence supports these theories, but also exposes where they are incomplete, most urgently the Standard Model does not explain the nature of dark matter, which plays an essential role in cosmology.
AI has the potential to answer these questions and others in physics.
For many physics problems, the governing equations encoding the fundamental laws of physics are well known. However, key calculations performed within these frameworks, as essential to test our understanding of the universe and guide the discovery of physics, may require computation or even be unreliable. IAIFI researchers develop AI for such first-principle theory studies, which require natural AI approaches that strictly encode physics knowledge.
“My group is developing new proven accurate algorithms for theoretical nuclear physics,” says Phiala Shanahan, an assistant professor of physics and LNS researcher at MIT. “Our approach to first principles appears to have applications in other areas of science and even in robotics, leading to exciting collaborations with sector partners.”
Incorporating physics principles into AI can also have a major impact on many experimental applications, such as designing AI methods that are easier to control. IAIFI researchers are working to improve the scientific potential of several facilities, including the Large Hadron Collider (LHC) and the Laser Interferometer Gravity Wave Observatory (LIGO).
“Gravity wave detectors are among the most sensitive instruments on earth, but the computing systems used to operate them are mostly based on technology from the last century,” said Lisa Barsotti, a scientist at the MIT Kavli Institute of Astrophysics and Space Research. “We are just beginning to scratch the surface of what can be done with AI; just enough to see the IAIFI change a game. ”
The unique features of these physics applications also offer broader research capabilities in AI broader. For example, physics-informed architecture and hardware development may lead to advances in the speed of AI algorithms, and work in statistical physics provides a theoretical basis for understanding AI dynamics.
“Physics has inspired many time-honored ideas in machine learning: maximizing entropy, Boltzmann machines, and variation tracking, to name a few,” says Pulkit Agrawal, an assistant professor of electrical engineering and computer science at MIT and a researcher in the Computer Laboratory for Science and Artificial Intelligence (CSAIL). “We believe that close interaction between physics and AI researchers will be the catalyst leading to the next generation of machine learning algorithms.”
Cultivating talent in the early career
AI technologies are advancing rapidly, making it both important and challenging to train junior researchers at the intersection of physics and AI. The IAIFI aims to recruit and train a talented and diverse group of early career researchers, including at the postdoc level through its IAIFI Fellows program.
“By offering our supporters their choice for research issues, and the opportunity to focus on emerging challenges in physics and AI, we will prepare many talented young scientists for future leaders in both academia and industry,” said MIT professor of physics Marin Soljacic of the Research Laboratory of Electronics (RLE).
IAIFI researchers hope that these sponsors will collaborate on interdisciplinary and multi-disciplinary activities, generate new ideas and approaches, translate physical challenges beyond their native domains, and help develop a common language across disciplines. Applications for the inaugural IAIFI supporters should appear by mid-October.
Another related effort spearheaded by Thaler, Williams, and Alexander Rakhlin, a university professor of brain and cognitive science at MIT and researcher at the Institute for Data, Systems, and Society (IDSS), is the development of a new interdisciplinary PhD program in physics, statistics, and data science, a collaboration between the Department of Physics and the Statistics and Data Science Center.
“Statistics and data science are among the fundamental pillars of AI. “Physics participating in the interdisciplinary doctoral program will create new ideas and areas of exploration, while promoting a new generation of leaders at the intersection of physics, statistics and AI,” says Rakhlin.
Education, broadcast, and collaborations
The IAIFI aims to cultivate “human intelligence” by promoting education and reach. For example, IAIFI members will contribute to the establishment of a MicroMasters study program at MIT for students from non-traditional backgrounds.
“We will increase the number of students in both physics and AI from under-represented groups by providing funding for the MicroMasters program,” said Isaac Chuang, professor of physics and electrical engineering, senior associate dean for digital learning, and RLE researcher at MIT. “We also plan to work with undergraduate MIT Summer Research Program students, to familiarize them with the tools of physics and AI research that they may not have access to at their home institutions.”
The IAIFI plans to expand its influence through numerous efforts, such as a K-12 program in which students receive data from the LHC and LIGO and are tasked with discovering the Higgs boson and gravitational waves.
“After confirming these recent Nobel Prizes, we can ask students to find small artificial signals embedded in the data using AI and fundamental physics principles,” said assistant physics professor Phil Harris, an LNS researcher at MIT. “With projects like this, we hope to spread knowledge about – and enthusiasm for – physics, AI and its intersection.”
In addition, the IAIFI will work with industry and government to push the boundaries of both AI and physics, as well as social sectors that benefit from AI innovation. IAIFI members already have many active collaborative partnerships with industry partners, including DeepMind, Microsoft Research, and Amazon.
“We will address two of the biggest mysteries of science: how our universe works and how intelligence works,” said MIT professor of physics Max Tegmark, a researcher at the MIT Kavli Institute. “Our key strategy is to link them, use physics to improve AI and AI to improve physics. We are pleased that the NSF is investing the vital seed funding needed to launch this exciting effort.”
Build new connections at MIT and beyond
With the help of MIT’s collaborative culture, the IAIFI aims to generate new connections and strengthen existing ones across MIT and beyond.
Of the 27 current IAIFI senior researchers, 16 are at MIT and members of the LNS, RLE, MIT Kavli Institute, CSAIL, and IDSS. In addition, IAIFI researchers are members of related NSF support efforts at MIT, such as the Center for Brains, Minds, and Machines within the McGovern Institute for Brain Research and the MIT-Harvard Center for Ultracold Atoms.
“We expect a lot of creative synergies when we bring physics and computer science together to study AI,” said Bill Freeman, the Thomas and Gerd Perkins professor of electrical engineering and computer science and researcher at CSAIL. “I’m happy to work with my physics colleagues on topics that span these fields.”
More generally, the IAIFI aims to make Cambridge, Massachusetts, and the surrounding Boston area a hub for collaborative efforts to advance both physics and AI.
“As we learn in 8.01 and 8.02, part of what makes physics so powerful is that it provides a universal language that can be applied to a wide range of scientific problems,” says Thaler. “Through the IAIFI, we will create a common language that crosses the intellectual boundaries between physics and AI to facilitate groundbreaking discoveries.”
.
