The Role of Foundational Research
A multi-organizational team including UT Austin, the University of Washington, Wichita State University, and Microsoft Research has been selected by the NSF as the nation’s designated Institute for Foundations of Machine Learning. The Institute is headquartered at UT’s newly launched Machine Learning Laboratory.
We are believers in the power of foundational research and its importance to both short and long-term innovation in AI. Most AI applications in use today are still relying on algorithms that were developed years ago, but the pace of modern technology is rapidly outstripping those algorithms’ capabilities.
Our researchers will develop new algorithms that can help machines learn on the fly—to change their expectations as they encounter people and objects in real life, and even to bounce back from deliberate attempts by adversaries to manipulate datasets.
IFML at NeurIPS
IFML authors have a strong presence at the Thirty-Fourth Conference on Neural Information Processing Systems. read more
Advanced Algorithms for Deep Learning
We will create fast, provably efficient tools for training neural networks and searching parameter spaces. We will develop new theories to rigorously explain successful heuristics.
Learning With Dynamic Data
Since datasets are constantly evolving, we must find new algorithms and models that can incorporate context and changes at training and test time, including robustness to perturbations.
Exploiting Structure in Data
What characteristics of a dataset help with training and inference? We will define and uncover rich mathematical structures in datasets in order to improve downstream modeling and optimization.
Optimizing Real-World Objectives
We will develop principled methods for automatically satisfying complex constraints and handling interactive feedback from users in real-world situations (e.g. safe robot navigation).
Use-Inspired Applications
To illustrate the impact of the above research themes, we have selected three use-inspired research areas: video, imaging, and navigation. We will work with Youtube, Netflix, and Facebook to establish a system of next-generation recognition and compression tools for video; we will partner with Dell Medical School to create fast, robust methods for MRI using novel priors; with Dell Technologies we will rethink noisy imaging problems for circuit design and integrity; and with the City of Austin and the mayor’s office, we will build modern systems to mitigate traffic congestion and rethink highway design.
