University of Cambridge > Talks.cam > Engineering Safe AI > An introduction to adversarial attacks and defences

  Browse A-Z

Log in

University Account External (via Google)

🔐 Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹ️ Information

An introduction to adversarial attacks and defences

Download to your calendar using vCal

If you have a question about this talk, please contact Adrià Garriga Alonso .

AI safety is not limited to RL settings. For example, we can use machine learning algorithms to design spam filters, yet attackers can still “reverse-engineer” our defence to send us junk emails. Autonomous driving systems based on computer vision techniques are also vulnerable to attacks, for instance, attackers can carefully apply a sticker to a stop sign in order to fool the vision system of the car. In this talk I will briefly discuss the mathematical framework of these attack techniques (specifically on image classifiers) and defence techniques against them.

Slides available here: http://yingzhenli.net/home/pdf/attack_defence.pdf

This talk is part of the Engineering Safe AI series.

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Buckingham House Lecture Theatre, Murray Edwards College The Divinity School, St. John’s College, University of Cambridge Cambridge University Engineering Dept., CBL Seminar Room (4th floor)

Other talks

Paediatric Medicine and Rheumatology Grand Rounds - Equine Reproducibility In Astrophysical Simulations: The AGORA Initiative and a 14-Year Human Experiment (In Memory of Joel Primack) Talk title TBC Optimal metric and approximation of free boundary minimal surfaces Title to be confirmed
 

© 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity