Our CVPR'21 paper on two novel model stealing attacks: (1) MAZE for completely data-free settings (no datasets needed) and (2) MAZE-PD when limited in-distribution data available. With Sanjay Kariyappa Sanjay Kariyappa and Moin Qureshi @mointweets. openaccess.thecvf.com/content/CVPR20…

Congratulations Akshitha. Hope you do great at CMU as a faculty member.

Sampoorna is my niece and a 10th grader. Very nice presentation on quantum superposition! Congrats Sampoorna! All the hallmarks of a future Professor!

Inspiring stories. Starting freshman year with no coding experience – New to CS youtu.be/-3z-xQ_ydFM via YouTube

The stop sign from our CVPR18 and WOOT18 paper on physical adversarial examples is on display at @DHMDresden AI exhibit! w/ earlence, @ivanevti, Tadayoshi (Yoshi) Kohno, Dawn Song, Atul Prakash, Florian Tramèr, Kevin Eykholt, Chaowei Xiao, Bo Li arxiv.org/abs/1707.08945

(1) Latest paper on our efforts at building least-privilege distributed computer systems, led by my student Yunang Chen will appear at USENIX Security 22! (Co-advised with Rahul Chatterjee). We show how to minimize sensitive data flowing into cloud IoT systems

A friend Nandit Soparkar’s interesting startup based on data they have collected over many years: You can predict and diagnose car problems with this Ann Arbor company's technology secondwavemedia.com/concentrate/in…

linkedin.com/posts/universi…

I am happy to announce that after 18 great years at Penn State I will be moving to the University of Wisconsin-Madison (UW–Madison), joining as the Tsun-Ming Shih Professor in the Dept. of Computer Sciences (UW–Madison Computer Sciences).

Attending my first conference today IEEE European Symposium on Security and Privacy! Come see my presentation on “GRAPHITE: Generating Automatic Physical Examples for Machine-Learning Attacks on Computer Vision Systems” this Thursday, June 10th.

Euro S&P paper preprint on automatically discovering robust perturbation attacks. arxiv.org/abs/2002.07088

FB reminded me: 3 May 2016 It's been so long since this one. My first "top conference" paper, my first exposure to media coverage and my first best paper award! For some reason people still think I do IoT/Smart Home stuff (I don't), but it was my big break in computer security.

Can we interpret why your out-of-distribution (OOD) detector thinks an input is in-distribution or OOD in terms of human-understandable concepts? Then check out our new ICML2023 paper, "Concept-based Explanations for Out-Of-Distribution Detectors" arxiv.org/pdf/2203.02586…

Chandrayaan-3 Mission: Here are the first observations from the ChaSTE payload onboard Vikram Lander. ChaSTE (Chandra's Surface Thermophysical Experiment) measures the temperature profile of the lunar topsoil around the pole, to understand the thermal behaviour of the moon's

Detecting Social Engineering Scams While Preserving User Privacy in the Digital Era (Atul Prakash, Kumar, Elisa Tsai - Electrical & Computer Engineering at Michigan) #ConPro24 conpro24.ieee-security.org/papers/prakash…

We’re happy to announce that Amrita Roy Chowdhury will be joining our faculty as an assistant professor. Welcome to CSE! ▶️Learn more about Professor Chowdhury: sites.google.com/wisc.edu/amrit…

Congratulations @Maggiemakar on winning the 2024 Google Research Scholar Award for work titled "Reliability through causal alignment: causally motivated models for pain management." Causally-aligning ML models to make them trustworthy is critical in healthcare settings.

🎉Proud to announce! 16 papers from our brilliant CSE researchers will be showcased at #SODA2025 happening Jan 12-15 in New Orleans! 🚀 #GraphTheory #Optimization #DataStructures #UMCSE 👏Congrats to all 18 CSE authors! 👇More: cse.engin.umich.edu/stories/sixtee…

We’re happy to announce that Travis Doom will be joining our teaching faculty as a Lecturer III in Fall 2025. 💫 Welcome to CSE! Travis Doom #UMCSE 👉Learn more about Travis: people.wright.edu/travis.doom

See our new work, FOCAL, at ICML 2025 for adding robust perception at test-time to vision foundation models such as CLIP under lighting shifts, rotations, etc. without retraining the foundation model. Congrats Ryan Feng Utkarsh Singhal @ NeurIPS and Stella Yu.