Deep Learning & Real World Experiences

Were back at Inspire9 for what promises to be another great evening...

Please help clear the tables set up the chairs if you arrive early. The chairs are located just by the lift. If you see the pizza guys outside then please give them a hand. As usual, help yourself to pizza and beer and soft drink in the beer fridge. When we are done, please help tidy up put the chairs away - every little helps!

Schedule:

6:00pm arrive and have pizza and beer
6:25pm introductions
6:35pm Zhen's talk
7:20pm Q & A for Zhen
7:30pm 10 min break
7:40pm Patrick's talk
8:30pm Q & A for Patrick
9:00pm Continue debate @ Post Office Hotel on Bridge Road

Zhen He: Recent Advances in the Field of Deep Learning

The area of neural networks has undergone revolutionary changes in the past 4 years. This has resulted in moving neural networks from the fringes of the machine learning community to center stage. Recently researchers have discovered effective ways of training neural networks that are 10 layers deep that contain billions of parameters. These deep neural networks have produced outstanding results in the area of image recognition, speech recognition, natural language processing, multi-modal learning, reinforcement learning, etc. For example in the area of image recognition the state-of-the-art deep learning system can achieve an error rate under 5%, which is more accurate than most humans can achieve.

In this talk I will present some recent applications of deep learning to solve very interesting problems. I will explain how deep learning works in an easy and approachable manner. I will present programming tools for implementing deep learning solutions to solve machine learning problems. I will also present tips on how to train deep learning algorithms to achieve high accuracy.

Zhen's Biography:

Dr Zhen He is an Associate Professor in the department of computer science and computer engineering in La Trobe University. He received his undergraduate and PhD degrees in the area of computer science from the Australian National University. His main research interest is in large scale machine learning. In particular he and his research team are exploring various ways of using the Apache Spark Big Data framework to scale out machine learning algorithms. A current focus is on scaling out deep learning algorithms using the Apache Spark framework.

Patrick Hall: Real World Machine Learning Experiences

We all know that Silicon Valley tech. firms are using machine learning, but what about the broader economy around the world? Are governments and companies in more traditional verticals like finance, healthcare, manufacturing, and oil production able to turn the ultra-hyped opportunities presented by machine learning into real-life value? In this talk, I will discuss my experiences with SAS customers in the U.S., Europe, Asia, and Australia as they seek to incorporate distributed computing, machine learning, and deep learning into their production data infrastructures. I will also point out some SAS resources for machine learning and deep learning, and how to integrate SAS with your favorite open-source packages. For a sneak peek, take a look here:

https://github.com/sassoftware/enlighten-apply
https://github.com/sassoftware/enlighten-integration (https://github.com/sassoftware/enlighten-deep)
https://github.com/sassoftware/enlighten-deep

Patrick's Biography:

Patrick designs new data mining and machine learning technologies for SAS. He is a Cloudera certified data scientist and a certified SAS Enterprise Miner predictive modeler. Patrick is also a co-author of two recent patent applications regarding unsupervised learning. He studied computational chemistry at the University of Illinois before graduating from the Institute for Advanced Analytics at North Carolina State University