MIT method keeps machine learning models of chaotic systems within physical bounds
Researchers at MIT's Laboratory for Information and Decision Systems have found a way to build physics-based limits into machine learning models of chaotic systems, such as weather, climate and turbulent flows, so that the models never drift into physically impossible predictions, the university says. The work is published in the Proceedings of the National Academy of Sciences by graduate student Sunbochen Tang and professors Navid Azizan and Themistoklis Sapsis.
Models can match data over short periods and still fail when run for a long time, as small errors accumulate until simulated values grow without limit, Azizan says. The team's approach lets a model learn complex dynamics while guaranteeing that its trajectories stay bounded, and MIT calls it the first work to establish such formal guarantees for machine learning models of chaos. The team is making its information freely available.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates
Industry analysis
2025 Global Business Services Agenda: Gen AI Takes Center Stage
This industry analysis by The Hackett Group explores the transformative impact of generative artificial intelligence (Gen AI) on global business services (GBS) in 2025. The study highlights the shift from exploration to acceleration of Gen AI initiatives, with 89% of executives advancing these projects to improve customer satisfaction, innovate products, and reduce costs. The report also discusses the challenges and strategies for successful Gen AI adoption, emphasizing the need for a technology-enabled operating model and the importance of reskilling the workforce.
Read moreYou may also like
PSI says a non-public Anthropic model helped prove a quantum computing conjecture open since 2008
OpenAI Says Its AI Model Refuted the Exponential Time Hypothesis
METR says evidence cannot rule out drastic AI R&D acceleration within a couple of years
MIT's AI copilot lets engineers steer algorithmic structure design as it runs
MBZUAI's Institute of Foundation Models releases K2-Horizon, six fully open models up to 375B parameters
Daily AI Brief: the AI news that matters, in your inbox.