Chhavi Yadav ✈️ ICML
@chhaviyadav_
Trustworthy and Reliable AI | ZKPs | PhD from @ucsd_cse | @trustworthy_ml
Upon graduation, I paused to reflect on what my PhD had truly taught me. Was it just how to write papers, respond to brutal reviewer comments, and survive without much sleep? Or did it leave a deeper imprint on me — beyond the metrics and milestones? Turns out, it did. A…
Exciting work on ‘proactive’ membership inference! Check it out to know how to make your material attribution-ready with STAMPing!
At #ICML2025, I am super excited to introduce STAMP. This is a marriage b/w dataset inference & watermarking that finally(!) lets creators PROVE their content was used to train LLMs🔍 Its a MAJOR push taking the academic problem into real world. w/Saksham Rastogi @danish037 🧵
Catch my talk today at UAI's SafeAI Workshop, at 1pm EST! I will talk about how explanations can be operationalized in adversarial contexts and about how data valuation is a classic case of misaligned incentives :)
Had an amazing time @NewInML @icmlconf giving a talk on "What I Wish I knew before starting a PhD (but learnt the hard way)"! Loved the post-talk discussions and the heart warming messages :) Sharing slides since some people asked, link in the tweet below 👇
Very nice talk, especially if you are new to publishing!
Had an amazing time @NewInML @icmlconf giving a talk on "What I Wish I knew before starting a PhD (but learnt the hard way)"! Loved the post-talk discussions and the heart warming messages :) Sharing slides since some people asked, link in the tweet below 👇
What a gold mine.. Excellent lectures on RL of/for LLMs by the very amazing @ErnestRyu !
Lecture recordings for **Reinforcement Learning of Large Language Models** have been completed! The videos and slides for all of the chapters have been uploaded. (1/3)
Exciting work showing known-answer detection (KAD) , a popular method to identify prompt injections, using an LLM to classify inputs as clean or contaminated, violates core security principles!
LLM-integrated applications and agents are vulnerable to prompt injection attacks, in which adversaries embed malicious instructions within seemingly benign user inputs to manipulate the LLM’s intended behavior. Recent defenses based on known-answer detection (KAD) have…
@kamalikac giving a super insightful talk on memorization of object segmentation models and LLMs. Takeaway: more recent models are memorizing less of their pre-training data! Room 223
Also very proud labmate moment 😊@kc_meehan
Proud advisor moment: @CaseyMeehan speaking at the Panel in the Memorization for Trustworthy Foundation Models Workshop at #ICML2025
Excellent talk by @chhaviyadav_ on ZKP and accountable AI!
Listen to my talk on ‘Accountable AI with ZKPs: Certifying Fairness and Explanations under model Confidentiality’ given @SimonsInstitute today! Link : m.youtube.com/watch?v=dH2yqw…
Spending tmrw in Berkeley! Would love to chat with people about AI, grad life or check out your favorite restaurant in Telegraph Ave! Please DM if you haven’t heard from me and would like to meet up :))
Listen to my talk on ‘Accountable AI with ZKPs: Certifying Fairness and Explanations under model Confidentiality’ given @SimonsInstitute today! Link : m.youtube.com/watch?v=dH2yqw…
Next week at the Simons Institute, a workshop on Proofs. Join us! simons.berkeley.edu/workshops/proo…
Wow what a masterstroke! Didn’t expect this! But welcome the strategy!
We're excited to announce a second physical location for NeurIPS 2025, in Mexico City. By expanding our physical locations, we hope to address concerns around skyrocketing attendance and difficulties in obtaining travel visas that some attendees have experienced in the past few…
We're excited to announce a second physical location for NeurIPS 2025, in Mexico City. By expanding our physical locations, we hope to address concerns around skyrocketing attendance and difficulties in obtaining travel visas that some attendees have experienced in the past few…
I'll be presenting my poster today in the East Exhibition Hall A-B #E-701 from 11am-1:30pm & then head to @NewInML to give a talk at 1:55pm in West Meeting Room 211-214! Come say hi, can't wait for the exciting discussions!! :))
