Michael O'Neill

I am currently a PhD student at the University of Wisconsin-Madison. I am working with Professor Stephen Wright on continuous optimization theory, largely focused on the worst-case complexity of nonconvex optimization methods.

I completed my undergraduate degree in Computer Science and Mathematics at the Rensselaer Polytechnic Institute in 2014, and my master's degree in Computer Sciences at UW-Madison in 2016.

For a take away, check out my Resume/CV.

Some recent news and travel:

- Our recent paper "A Line-Search Descent Algorithm for Strict Saddle Functions with Complexity Guarantees" was accepted at the IMCL 2020 Workshop on "Beyond First Order Methods in ML Systems". You can find out more by checking out our 3 minute video created for the workshop.
- Was scheduled to present work on strict saddle functions at the 2020 INFORMS Optimization Society Conference at Clemson University in Greenville, South Carolina. Unfortunately, the event was postponed indefinitely.
- I presented my work on worst-case complexity for bound constrained optimization at the 2019 INFORMS Annual Meeting in Seattle, Washington.

Contact Me.

---

Research Publications and Preprintes

A Line-Search Descent Algorithm for Strict Saddle Functions with Complexity Guarantees.
Michael O'Neill, Stephen J. Wright.
ICML 2020 Workshop on "Beyond First Order Methods in ML Systems" (also in submission to Journal of Machine Learning Research)

A Log-Barrier Newton-CG Method for Bound Constrained Optimization with Complexity Guarantees.
Michael O'Neill, Stephen J. Wright.
IMA Journal of Numerical Analysis
[Paper]

A Newton-CG Algorithm with Complexity Guarantees for Smooth Unconstrained Optimization.
Clément W. Royer, Michael O'Neill, Stephen J. Wright.
Mathematical Programming
[Paper]

Behavior of Accelerated Gradient Methods Near Critical Points of Nonconvex Functions
Michael O'Neill, Stephen J. Wright.
Mathematical Programming (series B).
[Paper]