The SIAM Conference on Mathematics of Data Science will take place during May 5-8, 2020 at the Hilton Cincinnati. This is the first such meeting hosted by the Society for Industrial and Applied Mathematics; it provides an opportunity for those working in computational and theoretical mathematics/statistics with a focus on data science to present their work in a forum dedicated specifically to discussing the mathematical implications of such work.
SigOpt Research is excited to attend in this meeting; we have, in the past, attended numerous SIAM meetings including CSE19, CSE17 and AN18 and always found them to provide an exciting exchange of ideas from a variety of perspectives. We expect this meeting to be no different and encourage our readers to join us in Cincinnati. During these 4 days, SigOpt is participating in several events.
Bayesian Methods in Science and Engineering
SigOpt Research and Uber AI are joining forces to organize a minisymposium on Bayesian methods in science and engineering. SigOpt research engineer Michael McCourt and Uber AI optimization expert Matthias Poloczek have organized 8 speakers from across industry and academia to speak on how they use Bayesian modeling to power their research. The session takes place during the afternoon of Wednesday, May 6 in Salon 1. More information is available at the links below:
More information will be provided here as we approach the conference.
Mathematical Issues of Machine Learning
Twitter machine learning engineer Ting Gao has organized a minisymposium to discuss applications of machine learning across various topics in scientific computing. SigOpt research engineer Michael McCourt is happy to be presenting at this session, which will also be a minor Illinois Institute of Technology reunion (Ting, Michael and fellow presenter Jinqiao Duan were all at IIT in the early 2010s). Details for his talk are below.
- Title: Bayesian Optimization for Model Calibration in Electroencephalography
- Time/Location: Wednesday, May 6, 2:15pm, Salon B/C
- Abstract: Electroencephalography (EEG) involves taking signals from a person’s scalp and reconstructing the location of the electrical dipoles which generated those signals. We present an adaptation of Bayesian optimization applied to processing signals during an EEG scan. Doing so provides a sample-efficient strategy for analyzing brain signals while minimizing the computational cost associated with solving the associated PDE (which defines the forward component of this inverse problem).
SIAM meetings are the premier locations for recruiting top mathematical talent, and, as such, SigOpt will be attending the career fair. This is scheduled to take place nearly all day on Thursday, May 7 in the Rosewood Room. If you are interested in an position at SigOpt (full time or intern) in San Francisco, please stop by and say hi to us. If you cannot attend the career fair, reach out to us at [email protected]