Chaowei Xiao
@ChaoweiX
Followers
2K
Following
125
Media
54
Statuses
295
Assistant Professor @University of Wisconsin, Madison| Researcher@NVIDIA| Researcher on AI Safety/Security
Joined October 2020
I will be at CVPR from 10-12 and introduce our recent work on AI safety/security at Robust Foundation Model workshop Please feel free to reach out if you are interested in safey/security topic.
0
0
3
Access control is a key concept for the computer security domain to ensures only authorized users can access sensitive assets. In our ACL paper, we applied this classic security concept to the large language models domain for safety. #safety #LLM #acl2025.
🚨 New paper accepted to #ACL2025!.We propose SudoLM, a framework that lets LLMs learn access control over parametric knowledge. Rather than blocking everyone from sensitive knowledge, SudoLM grants access to authorized users only. Paper: 🧵[1/6]👇
0
0
5
RT @fwang_nlp: 🎉 Excited to share that our paper, "MuirBench: A Comprehensive Benchmark for Robust Multi-image Understanding", will be pres….
0
17
0
Please see our work in Hall 3 + Hall 2B #602 Friday 10 am. Zhengyue will be there to present our work.
0
1
5
RT @zhenjun_zhao: DreamDrive: Generative 4D Scene Modeling from Street View Images. @PointsCoder, @Boyiliee, @iamborisi, @Yuxiao_Chen_, Yan….
0
5
0
RT @drmapavone: Introducing DreamDrive, which combines the complementary strengths of generative AI (video diffusion) and neural reconstruc….
0
44
0
RT @ZRChen_AISafety: 🧐 Are LLM agents safe and reliable? .😱 No, and they can be easily manipulated to crash your car, cash out your stocks,….
0
5
0
We will have a workshop on Responsibly Building the Next Generation of Multimodal Foundational Models with our amazing keynote and panel speakers @uiuc_aisecure @davidbau @FeiziSoheil @anikembhavi @Qdatalab @jasonbaldridge @furongh @liliyu_lili @lijuanWang
0
3
12
I will give a talk at SFU@NeurIPS. However, I can not attend it in-person this time.
Attending NeurIPS'24?.Please mark your calendar for our special event "SFU@NeurIPS 2024" 9 speakers from both academia & industry!.Only a 10-min walk from the convention center!.Let’s enjoy exciting talks and open discussions!
0
1
14
RT @arvindr_: A collaboration with @AnimaAnandkumar @ChaoweiX Shengchao Liu, @NVIDIA, @intel and many others at @argonne and @argonne_lcf !….
0
3
0
I will give keynote at CCS-LAMPS Oct 14 on AI security in the era of Foundation Models. Feel free to ping me if you are interested in this topic and want to talk at CCS @acm_ccs #CCS #LargeLanguageModel #security #jailbreak.
CCS-LAMPS 2024 is up, showcasing 11 accepted papers and two renowned keynote speakers specializing in AI safety and security. @ravenben @ChaoweiX This is a fantastic opportunity to engage in the field! .👉 Don’t miss out—visit to explore the full program!
0
1
17
Please notice this Environmental Injection Attack!!!.
The first study on privacy leakage of generalist web agents:. Generalist web agents have evolved rapidly and demonstrated remarkable potential in automating tasks on real websites, but their safety risks remain nearly unexplored. Given a webpage with malicious content injected,.
0
0
5
Please refer to our recent work with @xuefeng_du and @SharonYixuanLi for hallucination detection. We find that w/o requiring knowing the label information, utilizing the language model’s latent representations can capture information related to truthfulness.
🚀Excited to share our NeurIPS 2024 @NeurIPSConf spotlight HaloScope! 🎉 HaloScope is a new SOTA method that significantly improves hallucination detection for LLMs using unlabeled LLM generations 🧵#NeurIPS2024. Paper: w/ @ChaoweiX, @SharonYixuanLi
0
3
29
RT @xuefeng_du: 🚀Excited to share our NeurIPS 2024 @NeurIPSConf spotlight HaloScope! 🎉 HaloScope is a new SOTA method that significantly im….
0
18
0