University of Cambridge > Talks.cam > NLIP Seminar Series > Towards secure and efficient DNNs

Towards secure and efficient DNNs

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Andrew Caines.

On the efficiency side, I will show our recent progress on neural network compression methods that help to drastically reduce the size and the compute of a particular neural network. In addition, I will have a slight touch on the custom hardware accelerator that we’ve built for accelerating DNN inference. On the safety side, I will demonstrate the portability of adversarial samples and show an alternative detection method that rejects cheap adversarial samples with minimal computation cost.

This talk is part of the NLIP Seminar Series series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2019 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity