We are the NSF AI Institute for Foundations of Machine Learning (IFML)
Designated by the National Science Foundation (NSF) in 2020, IFML develops the key foundational tools for the next decade of AI innovation. Our institute comprises researchers from The University of Texas at Austin, University of Washington, Wichita State University, Microsoft Research. Stanford University, Santa Fe Institute, University of California, Los Angeles, University of California, Berkeley, California Institute of Technology, and Arizona State University.
Our researchers create new algorithms that can help machines learn on the fly, change their expectations as they encounter people and objects in real life, and even bounce back from deliberate attempts by adversaries to manipulate datasets.
Featured
Neurips 2024 Tutorial: Beyond Decoding: Meta-Generation Algorithms for Large Language Models
One of the most striking findings in modern research on large language models (LLMs) is that, given a model and dataset of sufficient scale, scaling up compute at training time leads to better final...
-
Article
Joint IFML/MPG Symposium at Simons Institute
Read More
-
Article
Students and Faculty Connect at the Machine Learning Lab Matching Event
Read More
-
Article
New IFML Framework for Diffusion Models to be included in several production pipelines by teams in Google
Read More
-
Article
IFML Diffusion Seminar Series: Tutorial on the Mathematical Foundations of Diffusion Models for Image Generation
Read More
-
Article
Turbocharging Protein Engineering with AI
Read More
-
Article
NLP Modules for High School
Read More
Upcoming Events and Workshops
-
December10
Neurips 2024 Tutorial: Beyond Decoding: Meta-Generation Algorithms for Large Language Models
Talk by
December13IFML Seminar: 12/13/24 - Safe and Informative Imaging via Conformal Prediction and Generative Models
Talk by Phil Schniter, Professor, Electrical and Computer Engineering The Ohio State University
Previously Recorded Talks
New & Noteworthy
-
Article
IFML research collaborations power next generation of multimodal AI with OpenCLIP and DataComp
OpenCLIP, the main text/image encoder in Stable Diffusion, is in the top 1% of all Python packages and has more than 50,000 git clones per day. DataComp, is the first rigorous benchmark for advancing multimodal dataset creation.
-
Article
Smoothed Analysis for Learning Concepts with Low Intrinsic Dimension earns Best Paper Award at COLT 2024!
Authored by IFML Director Adam Klivans, students Gautam Chandrasekaran, Konstantinos Stavropoulos, IFML postdoc Vasilis Kontonis, and former UT CS PhD Raghu Meka
-
Article
Now, Later, and Lasting: 10 Priorities for AI Research, Policy, and Practice
Shaping the future of artificial intelligence
-
-