Doorkeeper

AI Night: High-impact Deep Learning Applications

Thu, 30 Nov 2017 19:00 - 21:30 JST

Spaces Otemachi

Otemachi Building 1-6-1 Otemachi, Chiyoda-ku 100-0004 Tokyo

Register

Registration is closed

Get invited to future events

¥1,000 prepaid / ¥1,000 at the door
Food will be provided

Description

AI NIGHT: HIGH-IMPACT DEEP LEARNING APPLICATIONS

“AI is the new electricity” -- Andrew Ng
Artificial Intelligence is everywhere. We see the magic in things like image recognition, machine translation, robotics and predictive modeling. With the explosion of computational power and access to more data, Deep Learning is impacting industries and everyday life. This event aims to give high-level insights on new developments in the AI and Deep Learning space and facilitate Q&A and networking sessions.

Talk 1: Adam Gibson, CTO of Skymind
Unsupervised Deep Learning: Automatic Labeling For Time Series Data

Supervised Deep Learning has many successes, but generally depends on labeled data. However, labeled data is scarce and time consuming to obtain. Adam Gibson shows how to overcome those difficulties by using variational autoencoders to automatically label time series location data.

Adam Gibson is the CTO and co-founder of Skymind, a Deep Learning startup focused on enterprise solutions, and the co-author of “Deep Learning: A Practitioner’s Approach”.

Talk 2:

Erik Sjoberg, Senior Research Engineer, Ascent Robotics
Learning to Generalize: Leveraging Knowledge from Related Tasks

Although traditional deep networks are often trained from a random initialization, humans are able to leverage a wide variety of related experience to achieve good performance at new tasks with very few samples of training data. This talk will cover recent methods for learning to generalize from data and environments and continue to apply this knowledge as the target data distribution changes. These methods open the door to more robust systems which can continue to learn from experience.

Erik Sjoberg is a senior research engineer at Ascent Robotics, where he leads applied AI research and the development of robotic systems. Formerly an R&D systems engineer at Hewlett-Packard, Erik obtained a Masters in Robotic Systems Development from Carnegie Mellon.

This event is primarily designed for people who have some knowledge of Machine Learning and Deep Learning, but we welcome and encourage engineers and people from all experience levels to join.

Agenda
7:00 pm Doors open
7:30 pm Talk 1: Adam Gibson
8:00 pm Talk 2: Erik Sjoberg
8:30 pm Networking
Light snacks & drinks provided! We have one O’Reilly “Deep Learning: A Practitioner’s Approach” by Josh Patterson and Adam Gibson as a give away!

About this community

Dl4j users group

Dl4j users group

Come talk about dl4j usage in production as well as side projects! See more at deeplearning4j.org or come chat with our japanese community: https://gitter.im/deeplearning4j/deeplearning4j/deeplear...

Join community