Deep Learning Study: World Models

World Models is a brilliant adaptation of the environment modeling scenario within Reinforcement paradigm. Dr. David Ha discovered the following interesting insights from the games: Models of the environment using an unsupervised low level representation along with a recurrent memory units help the agent account for the next state in its decision making process. By training the policy and actions in the model of the environment, he was able to learn and prepare for behavior in the actual environment.

Here is the paper we will study this week:

David Ha et al, "World Models", May 2018
https://arxiv.org/abs/1803.10122

This discussion is co-led by Pushkar Merwah and Duane Nielsen.

Speakers' bio:
Pushkar Merwah is a founder and management strategy consultant specializing in corporate and AI strategies. He has been working on AI solutions pertaining to speech and visual data in a distributed edge environments. Prior to launching Left Brain LLC, Pushkar lead the Philips Lumileds operations strategy. He also worked at IBM for 11 years in various technical and strategy roles. He has a MBA from Columbia University & London Business School and an MS in Microelectronics Engineering from RIT.

Duane Nielsen is an Australian consultant, engineer and computer scientist. Duane has 15 years of experience in the software industry, including safety critical software engineering, satellite communications, web apps and application performance management. Duane has consulted with fortune 500 companies throughout North America and Asia Pacific, helping them scale-up application operations to service massive numbers of users.

Agenda:
6:30–7 pm Meet and greet
7-8 pm Presentation and discussion
8-8:30pm Social and catch up