JST CREST Edge-AI Team (Keio Univ, Univ Tokyo, Fixstars)

Exhibitor Information

"On-device Learning for Personalized Anomaly Detection Everywhere"

We will provide a solution of "anomaly detection everywhere" based on an on-device learning algorithm, associated technologies, and their chip integration for highly productive industries as well as a safe and secure society. Our solution eliminates the efforts on preparing accurate training data sets so that anyone can easily deploy it to anywhere.

- Personalized anomaly detection by on-site training without labeled training data
- Drastic cost reduction of neural network training that enables the on-site training on resource-limited edge devices
- Applications including factory, datacenter, home automation, UAV, surveillance camera, and so on

Contact Information

PR Contact Information

Links

URL 1

https://www.youtube.com/playlist?list=PL0RaY1M9ndKrafxpYkJhRUZh34GKU4qEM

On-device learning demo videos (Youtube)

URL 2

https://www.osluad.jp/

On-device learning demo site

URL 3

https://www.jst.go.jp/kisoken/crest/en/project/1111094/1111094_2017.html

JST AI-CREST projects (started in 2017)

URL 4

http://www.hal.ipc.i.u-tokyo.ac.jp/~kondo/

Kondo Lab., Graduate School of Information Science and Technology, The University of Tokyo

Event Information