Large-scale Deep Learning for AI

Hosted by

Co-Hosted by


Feb 18th, 4:30 - 6pm
(4:30-5pm networking, 5-6pm presentation with Q&A)


Cafeteria at Baidu USA,
1195 Bordeaux Dr. Sunnyvale California 94089
(onsite parking is free)

Event Fee

Free to all current SFASA members.
You are welcome to join the chapter on site.

  • Regular membership fee is $9
  • Student membership is $3 per year.
  • Please have a check of the amount due paid to SFASA and a photo ID ready for this purpose. If paying by cash, please have the exact amount.
  • We do not accept credit card payment at this moment


Pre-registration is required for this event. Please send an email to by February 16th 12pm PST with the following information:

  • Name, current affiliation and contact information
  • Are you a current member or plan to renew the membership/join the chapter on site

Large-scale Deep Learning for AI


Can machines think? That is a question raised by Alan Turing 66 years ago. A machine is regarded as intelligent when it can solve problems that human can solve. In recent years, deep learning has pushed significant performance improvement in many tasks, such as speech recognition and image understanding. The secret power of deep learning has been unleashed by big data, large scale computing infrastructure, and better algorithms.

In this talk, we will present some latest technology advances that tremendously enhance machines’ understanding of images, videos, and human language. A novel deep semantic network developed internally enables better parse intentions from user queries, which leads to more enhanced search results and more relevant contents. Empowered by deep recurrent networks, this system can generate human like responses in casual chat, and answering factoid questions from large knowledge bases. Combined with an in-house trained large-scale image classification network these newly created models are able to describe images and videos using human language, and to answer questions against objects in the images; all of which substantially enhance user experience in web-scale image search. These technologies enable us to build intelligent systems that better understand human behavior and better serving hundreds of millions of users.


Dr. Lei Li
Principal Research Scientist
Institute of Deep Learning, Baidu Research

Dr. Lei Li is Principal Research Scientist at Baidu's Institute of Deep Learning. His research interest lies in the intersection of machine learning, statistical inference, and natural language processing. He has published over 30 papers on Bayesian inference in open universe probabilistic models, probabilistic programming language, large-scale learning, time series, and social networks, and holds three US patents. He has served in the Program Committee for ICML 2014, ECML/PKDD 2014/2015, SDM 2013/2014, IJCAI 2011 / IJCAI2013 / IJCAI2016, KDD 2015/2016, and as a lecturer in 2014 summer school on Probabilistic Programming for Advancing Machine Learning. He has been invited as reviewer for TOMCCAP, DAMI, TKDE, TOSN, Neurocomputing, KDD, SIGMOD, VLDB, PKDD and WWW. He has been invited to review NSF proposal in 2010 and to DARPA's Information Science and Technology (ISAT) probabilistic programming workshop in 2013. Previously, he worked briefly at Microsoft Research (Asia and Redmond), Google (Mountain View), and IBM (TJ Watson Reserch Center).

Lei received his B.S. in Computer Science and Engineering from Shanghai Jiao Tong University in 2006 (ACM honored class) and Ph.D. in Computer Science from Carnegie Mellon University in 2011, respectively. His dissertation work on fast algorithms for mining co-evolving time series was awarded 2012 ACM KDD best dissertation (runner up). Before joining Baidu, he was working with Prof. Stuart Russell in EECS department of UC Berkeley as a Post-Doctoral Researcher.

Sponsored by SFASA and DahShu