[Colloquium] 12/4 Young Researcher Seminar Series: Surbhi Goel, University of Texas

[Alicia McClarin]

When:     Wednesday, December 4th, 2019 at 11:00am

Where:    TTIC, 6045 S Kenwood Avenue, 5th Floor, Room 526

*Who:*       Surbhi Goel, University of Texas


*Title: *Exploring Surrogate Losses for Learning Neural Networks


*Abstract: *Developing provably efficient algorithms for learning commonly
used neural network architectures continues to be a core challenge in
machine learning. The underlying difficulty arises from the highly
non-convex nature of the optimization problems posed by neural networks. In
this talk, I will discuss the power of convex surrogate losses for tackling
this underlying non-convexity. I will focus on the setting of ReLU
regression and show how the convex surrogate allows us to get approximate
guarantees in the challenging agnostic model. I will further show how these
techniques give positive results for simple convolutional and fully
connected architectures.



*Host:   *Nati Srebro <nati at ttic.edu>

-- 
*Alicia McClarin*
*Toyota Technological Institute at Chicago*
*6045 S. Kenwood Ave., **Office 504*
*Chicago, IL 60637*
*773-834-3321*
*www.ttic.edu* <http://www.ttic.edu/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20191127/a1e4ee25/attachment.html>