|Title:||Algorithmic and System Interface of Distributed Machine Learning|
|Date:||January 7, 2015 (Wednesday)|
|Time:||2:30 p.m. - 3:30 p.m.|
|Venue:||Room 1027, 10/F, Ho Sin-hang Engineering Building,
The Chinese University of Hong Kong,
|Speaker:||Dr. Qirong Ho
In many modern applications built on massive data and using high-dimensional models, such as web-scale content extraction via topic models, genome-wide association mapping via sparse regression, and image understanding via deep neural network, one needs to handle Big Machine Learning problems that challenge the limits of current infrastructures and algorithms. Although there have been point solutions to specific ML problems and applications (usually at a high engineering cost that is beyond the reach of most practitioners), a truly systematic solution to the Big ML challenge requires an algorithmic and systems interface that makes it possible for the average practitioner to build scalable ML programs. In this talk, I will provide a high-level overview of the algorithmic interface to Big ML --- the parallel workhorse optimization and MCMC algorithms that can be generically applied to many ML problems --- as well as the systems interface to Big ML --- the software frameworks that enable said workhorse algorithms to be easily deployed over a compute cluster. By covering the algorithmic and systems interface to Big ML, this talk will expose the many opportunities that lie within the space of Big ML, and serve as an invitation for statisticians, machine learners and computer scientists to discuss and collaborate on the interdisciplinary challenges that need to be solved.
Qirong Ho is a scientist at the Institute for Infocomm Research, A*STAR Singapore, and an adjunct assistant professor at the Singapore Management University School of Information Systems. His research focuses on two areas: statistical models for large-scale social media analysis, particularly latent space models for network visualization, community detection, user personalization and interest prediction. He also works on general-purpose software systems to enable large-scale distributed Machine Learning, such as the Petuum project (www.petuum.org) to accelerate the execution of large-scale Machine Learning algorithms based on their unique properties.
Enquiries: Miss Evelyn Lee at tel 3943 8444
For more information, please refer to http://www.cse.cuhk.edu.hk/seminar.