What to look forward to
The talk will explore how natural structure in data and learning tasks can enable efficient algorithms for problems believed to be computationally hard, including linear classification, sparse regression, and graphical-model learning. It will highlight key results and broader directions for understanding when structure can overcome computational barriers.
Inside the event
| Algorithmic innovation has been an important driver of progress in machine learning. However, several fundamental learning problems have resisted efficient algorithms, and there is evidence that they are computationally hard. My work finds ways around these barriers by identifying natural structure in the data or in what is being learned, and designing algorithms that exploit it. In this talk, I will discuss several results illustrating this perspective, drawing on problems ranging from linear classification and sparse linear regression to learning graphical models. I will highlight the different kinds of structure that make efficient learning possible, and discuss broader directions toward understanding when natural structure can overcome computational barriers.
Meet in Allston, MA
City view. The source has not supplied a venue; check the organizer for the exact location.
Open in Google Maps ↗Details come from the event listing; registration is managed by the organizer.
Last checked: Nov 7, 2026. Confirm current availability before making plans.