Zhuoran Liu

📢🚨 AICrypt 2024 will be held on May 26 in Zurich, Switzerland, as an affiliated event of @IACReurocrypt. Extended abstracts of works at the intersection of #AI and #cryptography are welcome. See the call at: aicrypt2024.aisylab.com Submission deadline: April 5, 2024 AoE

Are you interested in speech technology and (adversarial) machine learning? Apply to do your PhD in the area of speech processing for data protection @Radboud_Uni in beautiful Nijmegen, Netherlands: ru.nl/en/working-at/…

I will present our paper (led by @liuzrcc) @icmlconf. We find that adversarial training isn't necessary for mitigating (non-adaptive) data poisons, but simple image pre-processing (PP) is sufficient. In particular, which PP to choose depends on the perturbation frequencies.

if you would like to work on/are working on/are revisiting transferable adversarial attacks in ML, our paper will be a great start for you with simple idea, strong performance and public code.

One of the greatest moments as a Ph.D. student is to see your paper published online 🌹🙌 . PerBlur is finally online😀. A work by me, Alan Hanjalic (@alanhanjalic), and Martha Larson (@ngrams ). link: sciencedirect.com/science/articl… (1/N)

Code at github.com/liuzrcc/AIP You can find the presentation video at youtube.com/watch?v=cKXdDz… #TheWebConf #WWW21

Today at 10.00. Our work with @ngrams: "Adversarial Item Promotion: Vulnerabilities at the Core of Top-N Recommenders that Use Images to Address Cold Start" will be presented at @TheWebConf. Join us in room64 (Adversarial Cloaking and Attacks). Details at arxiv.org/abs/2006.01888