11 citations · 24 across the 11 of their papers we have counts for
12 papers
Inadvertent Context Leakage in Language Models
Jaiden Fairoze, Neal Mangaokar, Kamalika Chaudhuri +2
For AI agents to be useful beyond simple chat, they must hold sensitive user context such as calendars, credentials, health records, and financial data. We study whether the mere p…
Bypassing Prompt Guards in Production with Controlled-Release Prompting
Jaiden Fairoze, Sanjam Garg, Keewoo Lee +1
Ball et al. recently established that prompt filtering for AI alignment faces a fundamental barrier: under standard cryptographic assumptions, no filter running significantly faste…
Black-Box Crypto is Useless for Pseudorandom Codes
Sanjam Garg, Sam Gunn, Mingyuan Wang
A pseudorandom code is a keyed error-correction scheme with the property that any polynomial number of encodings appear random to any computationally bounded adversary. We show tha…
SoK: Watermarking for AI-Generated Content
Xuandong Zhao, Sam Gunn, Miranda Christ +11
As the outputs of generative AI (GenAI) techniques improve in quality, it becomes increasingly challenging to distinguish them from human-created content. Watermarking schemes are…
Publicly-Detectable Watermarking for Language Models
Jaiden Fairoze, Sanjam Garg, Somesh Jha +3
We present a publicly-detectable watermarking scheme for LMs: the detection algorithm contains no secret information, and it is executable by anyone. We embed a publicly-verifiable…
Overparameterization from Computational Constraints
Sanjam Garg, Somesh Jha, Saeed Mahloujifar +2
Overparameterized models with millions of parameters have been hugely successful. In this work, we ask: can the need for large models be, at least in part, due to the \emph{computa…
Deletion Inference, Reconstruction, and Compliance in Machine (Un)Learning
Ji Gao, Sanjam Garg, Mohammad Mahmoody +1
Privacy attacks on machine learning models aim to identify the data that is used to train such models. Such attacks, traditionally, are studied on static models that are trained on…
NeuraCrypt is not private
Nicholas Carlini, Sanjam Garg, Somesh Jha +3
NeuraCrypt (Yara et al. arXiv 2021) is an algorithm that converts a sensitive dataset to an encoded dataset so that (1) it is still possible to train machine learning models on the…
Is Private Learning Possible with Instance Encoding?
Nicholas Carlini, Samuel Deng, Sanjam Garg +6
A private machine learning algorithm hides as much as possible about its training data while still preserving accuracy. In this work, we study whether a non-private learning algori…
A Separation Result Between Data-oblivious and Data-aware Poisoning Attacks
Samuel Deng, Sanjam Garg, Somesh Jha +3
Poisoning attacks have emerged as a significant security threat to machine learning algorithms. It has been demonstrated that adversaries who make small changes to the training set…
Formalizing Data Deletion in the Context of the Right to be Forgotten
Sanjam Garg, Shafi Goldwasser, Prashant Nalini Vasudevan
The right of an individual to request the deletion of their personal data by an entity that might be storing it -- referred to as the right to be forgotten -- has been explicitly r…
Adversarially Robust Learning Could Leverage Computational Hardness
Sanjam Garg, Somesh Jha, Saeed Mahloujifar +1
Over recent years, devising classification algorithms that are robust to adversarial perturbations has emerged as a challenging problem. In particular, deep neural nets (DNNs) seem…