I'll be at ICML next week presenting PPT-Eval: A Benchmark for Computer-Use Agents on PowerPoint Tasks (https://t.co/euWVu2iPAM).
Come say hi! :)
This was a fun project with many friends at Microsoft on building nuanced rubrics for grading GUI-based agents on iterative slide/presentation editing tasks.
Can AI agents make and modify PowerPoint slides?
This new benchmark by Microsoft+us at CMU examines this, with rigorous evaluation. Take a look, and say hi to @apurvasgandhi at ICML if you're interested!
Think your mobile digital wallet is safe? Think again. This article dives into why that's not enough and how thieves can use your card even after you cancel it.
https://t.co/RjM2UXYM0J
@SyedRafiul#Cybersecurity#DigitalWallet#FinanceSecurity
Presenting our paper “Cloud Nine Connectivity: Security Analysis of In-Flight Wi-Fi Paywall Systems” today at #WiSec2025!
Huge thanks to my coauthors, esp. @RHasnainAnwar.
https://t.co/uTZ6ShfMQM
Do you know what your #digital wallet is up to? According to @umassengin researchers, you might not! They found critical #security loopholes that could threaten the safety of your #creditcard. How do we deal with this? More info: https://t.co/StVkCrFlyG
NDSS review: "this labeling scheme is a joke", "you’ve presenting results that claim it’s a solved problem. It’s irresponsible", "This isn’t an interesting problem and it’s not a useful result to know"
Me: can you calm down...
Honestly, such review hurts students a lot.