A Short History of and Introduction to Deep Learning

For our March event, we are thrilled to have John Kaufhold from Deep Learning Analytics present a technical introduction to Deep Learning, one of the hottest topics in data science in the last couple of years. How hot? Go search for "deep learning" (https://www.google.com/search?hl=en&q=deep+learning#hl=en&q=deep+learning&tbm=nws), and skim through hundreds of hyperventilating news articles describing how it's used at Google, Facebook, Netflix, and more, and how it's beating image and speech recognition benchmarks at near-human levels of performance. At it's core, Deep Learning is in many ways just the next iteration of the venerable Artificial Neural Network, a repeatedly hyped machine learning technique almost as old as the digital computer. So what's real innovation, what's hype, how do Deep Learning nets actually work, what's new about them, and what does it matter to you, the data science practitioner? Join us and find out!

NOTE: We are extremely grateful to Arlington Economic Development (http://www.arlingtonvirginiausa.com/), whose Arlington Meetup initiative helped us get access to an amazing venue for this event! We will be at the Artisphere (http://www.artisphere.com/), in Rosslyn. Plan to stick around for on-site Data Drinks afterwards!

6:30pm -- Networking, Food, and Refreshments

7:00pm -- Introduction

7:15pm -- Presentation and discussion

8:30pm -- Data Drinks (on-site cash bar!)

Abstract:

Big data and the emergence of data science as a formal discipline have both renewed interest in machine learning technologies that are scalable, fast, affordable and do not suffer from overfitting. Though the "No Free Lunch theorem" implies no machine learning technology in general can be expected to outperform all others on all tasks, some machine learning algorithms have been shown to consistently outperform others in empirical studies. For example, recent theoretical, algorithmic and practical breakthroughs in Deep Learning have been rapidly adopted and applied to industrial big data applications by the likes of Google, Apple and Facebook. Google+ image search and Siri, for example, both currently exploit Deep Learning algorithms developed in the past few years. In this talk I will discuss some recent Deep Learning history in the broader context of machine learning, highlighting the influence of Restricted Boltzman Machines, unsupervised feature learning, Dropout, rectified linear units, hierarchical distributed representations in deep architectures, GPU hardware acceleration, and open benchmarks."

Bio:

Dr. Kaufhold is a data scientist and managing partner of Deep Learning Analytics (http://www.deeplearninganalytics.com/), a data science company based in Arlington, VA. Prior to forming Deep Learning Analytics, Dr. Kaufhold investigated deep learning algorithms as a staff scientist at NIH. Prior to NIH, Dr. Kaufhold was a Technical Fellow at SAIC, serving as principal investigator or technical lead on a number of large government contracts funded by NIH, DARPA and IARPA, among others. Prior to joining SAIC, Dr. Kaufhold investigated machine learning algorithms for medical image analysis and image and video processing at GE's Global Research Center. Dr. Kaufhold earned his Ph.D. from Boston University's biomedical engineering department in 2001.

Sponsors:

This event is sponsored by Arlington Economic Development (http://www.arlingtonvirginiausa.com/)/Arlington Meetup, Cloudera (http://www.cloudera.com/), Statistics.com (http://bit.ly/12YljkP), SynglyphX (http://www.synglyphx.com/), IBM Analytics Solution Center (https://www.ibm.com/ascdc), and Elder Research (http://datamininglab.com/). Would you like to sponsor too? Please get in touch!