KC
Kamalika Chaudhuri
cs.LGcs.CRstat.MLcs.AIcs.CVcs.CLcs.DBmath.STstat.THcs.IT
On Valency
published · living versionsW_wgx8xep8·v1 · currentpublished
InferPriv: Evaluating Inferential Privacy Leakage in LLM Assistants
with Pengrun Huang, Yiwei Fu, Luoxi Meng
1 version
W_73pumvjw·v1published
Preprints & journals
134 papers in the corpus · 2009–2026Repeat-After-Me: Black-Box Adaptive Visual Prompt Injection2609.04533v2 · Sizhe Chen, Yu-Lin Tsai, Ivan Evtimov et al.2026 · 0 citationsarXiv
Inadvertent Context Leakage in Language Models2608.19857v1 · Jaiden Fairoze, Neal Mangaokar, Kamalika Chaudhuri et al.2026 · 0 citationsarXiv
Alignment Defends LLMs from Property Inference Attacks2606.10217v1 · Pengrun Huang, Chhavi Yadav, Ruihan Wu et al.2026 · 0 citationsarXiv
Agent Security is a Systems Problem2605.18991v2 · Mihai Christodorescu, Earlence Fernandes, Ashish Hooda et al.2026 · 0 citationsarXiv
Muse Spark Safety & Preparedness Report2606.12429v1 · Cristina Menghini, Peter Ney, Hamza Kwisaba et al.2026 · 0 citationsarXiv
Dataset Watermarking with Provable Black-Box Detection2605.06865v2 · Pengrun Huang, Kamalika Chaudhuri, Yu-Xiang Wang2026 · 0 citationsarXiv
DPrivBench: Benchmarking LLMs' Reasoning for Differential Privacy2604.15851v3 · Erchi Wang, Pengrun Huang, Eli Chien et al.2026 · 0 citationsarXiv
How Vulnerable Are AI Agents to Indirect Prompt Injections? Insights from a Large-Scale Public Competition2603.15714v1 · Mateusz Dziemian, Maxwell Lin, Xiaohan Fu et al.2026 · 0 citationsarXiv
Safety Alignment of LMs via Non-cooperative Games2512.20806v3 · Anselm Paulus, Ilia Kulikov, Brandon Amos et al.2025 · 0 citationsarXiv
Privacy Blur: Quantifying Privacy and Utility for Image Data Release2512.16086v1 · Saeed Mahloujifar, Narine Kokhlikyan, Chuan Guo et al.2025 · 0 citationsarXiv
Z0-Inf: Zeroth Order Approximation for Data Influence2510.11832v1 · Narine Kokhlikyan, Kamalika Chaudhuri, Saeed Mahloujifar2025 · 0 citationsarXiv
RL Is a Hammer and LLMs Are Nails: A Simple Reinforcement Learning Recipe for Strong Prompt Injection2510.04885v1 · Yuxin Wen, Arman Zharmagambetov, Ivan Evtimov et al.2025 · 0 citationsarXiv
Embodied AI Agents: Modeling the World2506.22355v3 · Pascale Fung, Yoram Bachrach, Asli Celikyilmaz et al.2025 · 9 citationsarXiv
Machine Learning with Privacy for Protected Attributes2506.19836v1 · Saeed Mahloujifar, Chuan Guo, G. Edward Suh et al.2025 · 0 citations2025 IEEE Symposium on Security and Privacy (SP)
Learning-Time Encoding Shapes Unlearning in LLMs2506.15076v1 · Ruihan Wu, Konstantin Garov, Kamalika Chaudhuri2025 · 0 citationsarXiv
Can We Infer Confidential Properties of Training Data from LLMs?2506.10364v4 · Pengrun Huang, Chhavi Yadav, Kamalika Chaudhuri et al.2025 · 0 citationsarXiv
AbstentionBench: Reasoning LLMs Fail on Unanswerable Questions2506.09038v1 · Polina Kirichenko, Mark Ibrahim, Kamalika Chaudhuri et al.2025 · 3 citationsarXiv
Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awareness2506.05735v4 · Rongzhe Wei, Peizhi Niu, Hans Hao-Hsun Hsu et al.2025 · 0 citationsarXiv
How much do language models memorize?2505.24832v3 · John X. Morris, Chawin Sitawarin, Chuan Guo et al.2025 · 3 citationsarXiv
WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks2504.18575v3 · Ivan Evtimov, Arman Zharmagambetov, Aaron Grattafiori et al.2025 · 5 citationsarXiv
Measuring Déjà vu Memorization Efficiently2504.05651v1 · Narine Kokhlikyan, Bargav Jayaraman, Florian Bordes et al.2025 · 0 citationsarXiv
AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents2503.09780v3 · Arman Zharmagambetov, Chuan Guo, Ivan Evtimov et al.2025 · 6 citationsarXiv
ExpProof : Operationalizing Explanations for Confidential Models with ZKPs2502.03773v4 · Chhavi Yadav, Evan Monroe Laufer, Dan Boneh et al.2025 · 1 citationarXiv
A Closer Look at the Learnability of Out-of-Distribution (OOD) Detection2501.08821v1 · Konstantin Garov, Kamalika Chaudhuri2025 · 0 citationsarXiv
Auditing -Differential Privacy in One Run2410.22235v1 · Saeed Mahloujifar, Luca Melis, Kamalika Chaudhuri2024 · 0 citationsarXiv
Distribution Learning with Valid Outputs Beyond the Worst-Case2410.16253v1 · Nick Rittler, Kamalika Chaudhuri2024 · 0 citationsarXiv
Evaluating Deep Unlearning in Large Language Models2410.15153v4 · Ruihan Wu, Chhavi Yadav, Russ Salakhutdinov et al.2024 · 1 citationarXiv
SecAlign: Defending Against Prompt Injection with Preference Optimization2410.05451v3 · Sizhe Chen, Arman Zharmagambetov, Saeed Mahloujifar et al.2024 · 30 citationsarXiv
Influence-based Attributions can be Manipulated2409.05208v4 · Chhavi Yadav, Ruihan Wu, Kamalika Chaudhuri2024 · 1 citationarXiv
On Differentially Private U Statistics2407.04945v1 · Kamalika Chaudhuri, Po-Ling Loh, Shourya Pandey et al.2024 · 1 citationarXiv
Beyond Discrepancy: A Closer Look at the Theory of Distribution Shift2405.19156v1 · Robi Bhattacharjee, Nick Rittler, Kamalika Chaudhuri2024 · 0 citationsarXiv
An Introduction to Vision-Language Modeling2405.17247v1 · Florian Bordes, Richard Yuanzhe Pang, Anurag Ajay et al.2024 · 47 citationsarXiv
Better Membership Inference Privacy Measurement through Discrepancy2405.15140v1 · Ruihan Wu, Pengrun Huang, Kamalika Chaudhuri2024 · 0 citationsarXiv
Metric Differential Privacy at the User-Level Via the Earth Mover's Distance2405.02665v3 · Jacob Imola, Amrita Roy Chowdhury, Kamalika Chaudhuri2024 · 1 citationarXiv
Uncertainty-Based Abstention in LLMs Improves Safety and Reduces Hallucinations2404.10960v1 · Christian Tomani, Kamalika Chaudhuri, Ivan Evtimov et al.2024 · 2 citationsarXiv
Guarantees of confidentiality via Hammersley-Chapman-Robbins bounds2404.02866v3 · Kamalika Chaudhuri, Chuan Guo, Laurens van der Maaten et al.2024 · 0 citationsTransactions on Machine Learning Research, article 2496: pp. 1-20, 2024
DP-RDM: Adapting Diffusion Models to Private Domains Without Fine-Tuning2403.14421v3 · Jonathan Lebensold, Maziar Sanjabi, Pietro Astolfi et al.2024 · 0 citationsarXiv
Privacy Amplification for the Gaussian Mechanism via Bounded Support2403.05598v1 · Shengyuan Hu, Saeed Mahloujifar, Virginia Smith et al.2024 · 0 citationsarXiv
Differentially Private Representation Learning via Image Captioning2403.02506v2 · Tom Sander, Yaodong Yu, Maziar Sanjabi et al.2024 · 3 citationsarXiv
FairProof : Confidential and Certifiable Fairness for Neural Networks2402.12572v2 · Chhavi Yadav, Amrita Roy Chowdhury, Dan Boneh et al.2024 · 2 citationsarXiv
Measuring Privacy Loss in Distributed Spatio-Temporal Data2402.11526v1 · Tatsuki Koga, Casey Meehan, Kamalika Chaudhuri2024 · 0 citationsarXiv
Déjà Vu Memorization in Vision-Language Models2402.02103v2 · Bargav Jayaraman, Chuan Guo, Kamalika Chaudhuri2024 · 1 citationarXiv
Effective pruning of web-scale datasets based on complexity of concept clusters2401.04578v2 · Amro Abbas, Evgenia Rusak, Kushal Tirumala et al.2024 · 0 citationsarXiv
Differentially Private Multi-Site Treatment Effect Estimation2310.06237v1 · Tatsuki Koga, Kamalika Chaudhuri, David Page2023 · 1 citationarXiv
Unified Uncertainty Calibration2310.01202v2 · Kamalika Chaudhuri, David Lopez-Paz2023 · 0 citationsarXiv
Large-Scale Public Data Improves Differentially Private Image Generation Quality2309.00008v1 · Ruihan Wu, Chuan Guo, Kamalika Chaudhuri2023 · 1 citationarXiv
ViP: A Differentially Private Foundation Model for Computer Vision2306.08842v2 · Yaodong Yu, Maziar Sanjabi, Yi Ma et al.2023 · 0 citationsarXiv
Agnostic Multi-Group Active Learning2306.01922v1 · Nick Rittler, Kamalika Chaudhuri2023 · 0 citationsarXiv
Data Redaction from Conditional Generative Models2305.11351v2 · Zhifeng Kong, Kamalika Chaudhuri2023 · 4 citationsarXiv
Do SSL Models Have Déjà Vu? A Case of Unintended Memorization in Self-supervised Learning2304.13850v3 · Casey Meehan, Florian Bordes, Pascal Vincent et al.2023 · 1 citationarXiv
Career total: 220 works. 134 are in this corpus.Showing the 50 most recent.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.