
Assistant Professor @ National Taiwan Univ. Prev.: @Google @GeorgiaTech @UofIllinois @Amazon @BellLabs #RegulatableAI
Joined November 2018
- Tweets381
- Following405
- Followers408
- Likes1.7K
Pinned Tweet
[BITS]
The CFP for the special issue of Machine Unlearning on BITS is out! Check it out here: itsoc.org/bits/calls-for-pap…
I'm glad to prepare this special issue with @VarunChandrase3 and @PanLi90769257. Please reach out to us if you have any questions!
Eli Chien retweeted
[🔥🔥NeurIPS workshop-InfPriv-2026] Recent alarm from banks told us that giving agents access to reusable private credentials is risky. Agent era needs new digital infrastructure, which motivates our KITA study -- keeping personal keys (e.g., password) outside LLM processes. 🔥🔥
🤖 Made with AI
We are hiring! National Taiwan University (NTU), AI Center of Research Excellence is a newly established AI center at NTU, where I am also affiliated. If you are looking for a position as an Open-Rank Professor/Research Fellow, feel free to take a look!
academicjobsonline.org/ajo/j…
1
7
702
If you have any questions, feel free to reach out to me or the CEO @hsuantienlin of our center!
1
65
Eli Chien retweeted
My dream had always been to prove beautiful theorems as the AI agents unloaded my dishwasher. Well, it went the wrong way... 😂
I will be at #COLM in SF from 10/6-10/9. Let me know if you’re around and want to chat about trustworthy AI, especially in inference time and agentic settings!
#NeurIPS results are out! Whether your paper is accepted or rejected, you still have a chance to present at InfPriv @ #NeurIPS2026, Sydney! Consider submitting your work via our NeurIPS fast-track with minimal effort, where the deadline is 9/25 AOE.
beyond-private-training.ai.s…
How do we push AI privacy beyond standard DP training? 🤔 Join our #NeurIPS2026 Workshop in Sydney to tackle the biggest open questions in privacy for AI at inference time and AI for privacy!
Learn more: beyond-private-training.ai.s…
InfPriv @ NeurIPS 2026, Sydney deadline is approaching (about 4 days left)! Submit your work by 9/7 AOE! Join our discussion on AI for privacy and privacy for AI!
How do we push AI privacy beyond standard DP training? 🤔 Join our #NeurIPS2026 Workshop in Sydney to tackle the biggest open questions in privacy for AI at inference time and AI for privacy!
Learn more: beyond-private-training.ai.s…
Eli Chien retweeted
People used to be able to impress and intimidate reviewers with complicated proofs. This will change. In the age of AI inscrutable proofs are cheap. It is understandable proofs that are valuable. Opaque complexity is now it is a sign of laziness or lack of insight.
How do we push AI privacy beyond standard DP training? 🤔 Join our #NeurIPS2026 Workshop in Sydney to tackle the biggest open questions in privacy for AI at inference time and AI for privacy!
Learn more: beyond-private-training.ai.s…
It's my pleasure to organize this workshop with @ruihan_w @kamalikac @yuxiangw_cs @niloofar_mire @erchi_w @JiachenWang97 @ahonkela !
We also have a list of great speakers: @lin_zinan @ChulinXie @MihaelaVDS ,Dinh Thai Hoang, and Saeed Mahloujifar!
Heading to Toronto for the exciting workshop on the future of unlearning, hosted by Nick (Hengrui) Jia, Anvith Thudi and @NicolasPapernot ! Looking forward to the fruitful discussions :)
cleverhans.io/events/unlearn…
5
870
Generating Differentially Private (DP) synthetic data from proprietary LLM APIs just got more efficient. Introducing MAPLE (Metadata Augmented Private Language Evolution). 🍁Accepted by #COLM2026. #GoogleResearch #NTU #UIUC #Meta
📄 Read it here: arxiv.org/abs/2603.19258
Thanks to all my great collaborators @huyuzheng1999, Ryan McKenna, Shanshan Wu, Zheng Xu, @KairouzPeter !
1
1
112
Eli Chien retweeted
(4/n)
New adversarial testing tool for LLMs by decomposing malicious intents into harmless prompts and aggregating the responses from the target.
w/ Rongzhe Wei, Peizhi Niu, Xinjie Shen, Tong Tu, Yifan Li, Ruihan Wu @chien_eli Olgica Milenkovic, Pan Li
👉linkedin.com/feed/update/urn…
I'll attend #ICML2026! This is the first conference since I started my new job :). Let me know if you want to chat about trustworthy AI, especially on privacy, safety, and copyright at inference time!
Eli Chien retweeted
It's time for the Theory Jobs spreadsheet for 2026! The theory CS community crowdsources who accepted which jobs every year.
Link in next post 1/2
Eli Chien retweeted
It's so cringe when real people I otherwise know and respect post obvious AI slop on social media, particularly when they're (supposedly) expressing their feelings. Authenticity is so rare and valuable these days, and it's sad to see people just cede it from the get-go
Eli Chien retweeted
Can frontier models reasoning about Differential Privacy?
We are excited to release DPrivBench, a benchmark curated by DP experts to evaluate LLMs’ reasoning ability on Differential Privacy. @pengrun_huang @chien_eli @omthkkr @kamalikac @yuxiangw_cs @ruihan_w
DPrivBench contains two complementary tracks:
Category 1 (preliminary): fundamental DP mechanism questions focused on sensitivity calculation and noise calibration.
Category 2 (advanced): research-level DP algorithms from the literature that require advanced, algorithm-specific mathematical reasoning.
Our findings are both promising and cautionary: frontier models are strong on Category 1, yet still struggle with research-level DP algorithms requiring nuanced reasoning about assumptions, privacy accounting, and algorithm-specific guarantees.
[ICML 2026] Happy to share that we have two papers accepted by ICML 2026!
DP Theory: arxiv.org/abs/2506.00158
LLM Jailbreak: arxiv.org/abs/2512.01353
Thanks to all collaborators: @wnchen1994 ,@PanLi90769257, Rongzhe, Peizhi, @Frilk3 ,Tony ,Yifan @ruihan_w ,@pinyuchenTW ,Olgica













