CATCH MURI

58 posts

CATCH MURI banner
CATCH MURI

CATCH MURI

@catch_muri

The Joint MURI-AUSMURI in Cybersecurity Assurance for Teams of Computers and Humans (CATCH)

USA and Australia เข้าร่วม Temmuz 2021
39 กำลังติดตาม43 ผู้ติดตาม
ทวีตที่ปักหมุด
CATCH MURI
CATCH MURI@catch_muri·
Hello world! We're the MURI for Cybersecurity Assurance for Teams of Computers and Humans (CATCH). Will post ground-breaking research from our group of researchers, and developments in the areas of robust human-AI teams and cybersecurity. We're also here: catch-muri.org
Somesh Jha@jhasomesh

Humbled and excited to receive MURI on the general topic of robust learning. PIs are from US (UW, UCSD, CMU, PSU) and Australia (University of Melbourne, Macquarie University, University of Newcastle). cto.mil/wp-content/upl…

English
1
0
0
0
CATCH MURI รีทวีตแล้ว
Somesh Jha
Somesh Jha@jhasomesh·
@christodorescu and I hosted a panel at the CRA snowbird conference. The title of the panel was: Security Risks of Generative AI: From Identification and Mitigation to Responsible Use Our amazing panelists were John Mitchell (Stanford University), Rebecca Wright (Barnard College), Matt Turek (DARPA)
English
1
2
4
435
CATCH MURI รีทวีตแล้ว
Kamalika Chaudhuri
Kamalika Chaudhuri@kamalikac·
Last year, we had so much fun at this workshop! Come join @chhaviyadav_ and the rest of us at @NeurIPSConf and send your best interpretability papers here.
Chhavi Yadav@chhaviyadav_

📰 Excited to be organizing a workshop on Interpretability @NeurIPSConf'24, called 'Interpretable AI : Past, Present and Future' Submit to our workshop for all things inherently interpretable! Submission ddl : 30 Aug 🔗 interpretable-ai-workshop.github.io Follow this account for updates!

English
0
4
14
3.4K
CATCH MURI รีทวีตแล้ว
Somesh Jha
Somesh Jha@jhasomesh·
I thought a lot about this... (Attack->fix->..)^* cycle is fundamental to the crypto/security community. In fact, a lot of cool technical results won't be possible unless we had that cycle. For example, lattice based crypto came after a long sequence of crypto schemes that were broken (maybe someone can put actual examples). Having the (attack->fix)^* is fundamental to the security/crypto field.
Gautam Kamath@thegautamkamath

*Must read* for anyone interested in ML security, by Nicholas Carlini. Attacks are the only way we know whether or not a purportedly secure system actually is. Moreover, I consider personal attacks like this unacceptable in my research communities. nicholas.carlini.com/writing/2024/w…

English
0
3
21
5K
CATCH MURI
CATCH MURI@catch_muri·
Excited to be kicking off a @catch_muri workshop at @CarnegieMellon SV campus. Looking forward to two days of great discussions on robust human-AI teaming.
CATCH MURI tweet media
English
0
1
3
398
CATCH MURI รีทวีตแล้ว
Xuanli He
Xuanli He@zodiacJRH·
🚨 New Paper! (arxiv.org/abs/2404.19597)🚨 We uncover significant vulnerabilities in Multilingual LLMs (MLLMs) (e.g., BLOOM, Llama2, Llama3, Gemma, and GPT-3.5-turbo) to cross-lingual transferable backdoor attacks. #AIsafety #LLMs #backdoors
Xuanli He tweet media
English
2
8
24
6.7K
CATCH MURI รีทวีตแล้ว
Somesh Jha
Somesh Jha@jhasomesh·
Excellent venue on a very important topic. Also, some very interesting keynote talks are being arranged by the organizers of this workshop at @IEEESSP
Mihai Christodorescu@christodorescu

(Second Call for Papers) Submit your work on the security for GenAI systems and applications. *Security Architectures for Generative AI (SAGAI'24)* is a new workshop at IEEE S&P this year. Full CFP: sites.google.com/view/sagai2024… Submission deadline: February 5, 2024

English
0
2
4
1.8K
CATCH MURI รีทวีตแล้ว
Amrita Roy Chowdhury
Amrita Roy Chowdhury@AmritaRoyChowd8·
I'm on the academic job market I'm a full-stack data privacy researcher I build systems that are 1)provably private 2)functionally-rich 3)compatible w/ real-world constraints I do this by exploring the synergy between cryptography & differential privacy, both in theory & practice
English
10
52
189
115K
CATCH MURI
CATCH MURI@catch_muri·
Congrats to @catch_muri co-lead @jhasomesh and his coauthors on this groundbreaking work!
Somesh Jha@jhasomesh

Our paper presents the first provable watermarking scheme for large language models (LLMS) with public detectability or verifiability: we use a private key for watermarking and a public key for watermark detection. Work with Jaiden, Sanjam, Mingyuan, Mohammad, and Saeed. Comments most welcome. We will update in the next version. eprint.iacr.org/2023/1661.pdf

English
0
1
1
218

ค้นพบ

@jhasomesh @christodorescu @chhaviyadav_ @NeurIPSConf @CarnegieMellon @kamalikac @asia_biega @mlsec