49 citations · 83 across the 16 of their papers we have counts for
18 papers
Surprise Masking: Private Data Filtering with Model Likelihoods
Riya Tyagi, Sahra Ghalebikesabi, Vinnie Monaco +2
Data useful for language model (LM) training can contain private or unwanted information that is at risk of being learned or memorized. We propose Surprise Masking (SM), a simple f…
Unmasking the Secret Code: Using Canaries to Show How Masked Context Controls Memorization in SFT
Avi Schwarzschild, Om Thakkar, Sahra Ghalebikesabi +2
Canaries, or synthetic training records containing known secrets, are widely used to assess whether language models reveal information from their training data. But how much do the…
Model Card for OpenAI Privacy Filter
Charles de Bourcy, Sahra Ghalebikesabi, Avi Schwarzschild +22
OpenAI Privacy Filter is a compact, bidirectional token-classification model for detecting and redacting personally identifiable information (PII) and secrets in unstructured text.…
Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
Gheorghe Comanici, Eric Bieber, Mike Schaekermann +3431
In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our…
Exploring the limits of strong membership inference attacks on large language models
Jamie Hayes, Ilia Shumailov, Christopher A. Choquette-Choo +13
State-of-the-art membership inference attacks (MIAs) typically require training many reference models, making it difficult to scale these attacks to large pre-trained language mode…
CI-Bench: Benchmarking Contextual Integrity of AI Assistants on Synthetic Data
Zhao Cheng, Diane Wan, Matthew Abueg +6
Advances in generative AI point towards a new era of personalized applications that perform diverse tasks on behalf of users. While general AI assistants have yet to fully emerge,…
Operationalizing Contextual Integrity in Privacy-Conscious Assistants
Sahra Ghalebikesabi, Eugene Bagdasaryan, Ren Yi +10
Advanced AI assistants combine frontier LLMs and tool access to autonomously perform complex tasks on behalf of users. While the helpfulness of such assistants can increase dramati…
AirGapAgent: Protecting Privacy-Conscious Conversational Agents
Eugene Bagdasarian, Ren Yi, Sahra Ghalebikesabi +5
The growing use of large language model (LLM)-based conversational agents to manage sensitive user data raises significant privacy concerns. While these agents excel at understandi…
Explainable AI for survival analysis: a median-SHAP approach
Lucile Ter-Minassian, Sahra Ghalebikesabi, Karla Diaz-Ordaz +1
With the adoption of machine learning into routine clinical practice comes the need for Explainable AI methods tailored to medical applications. Shapley values have sparked wide in…
Differentially Private Statistical Inference through -Divergence One Posterior Sampling
Jack Jewson, Sahra Ghalebikesabi, Chris Holmes
Differential privacy guarantees allow the results of a statistical analysis involving sensitive data to be released without compromising the privacy of any individual taking part.…
Challenges and Opportunities of Shapley values in a Clinical Context
Lucile Ter-Minassian, Sahra Ghalebikesabi, Karla Diaz-Ordaz +1
With the adoption of machine learning-based solutions in routine clinical practice, the need for reliable interpretability tools has become pressing. Shapley values provide local e…
A Rigorous Link between Deep Ensembles and (Variational) Bayesian Methods
Veit David Wild, Sahra Ghalebikesabi, Dino Sejdinovic +1
We establish the first mathematically rigorous link between Bayesian, variational Bayesian, and ensemble methods. A key step towards this it to reformulate the non-convex optimisat…
AIRIVA: A Deep Generative Model of Adaptive Immune Repertoires
Melanie F. Pradier, Niranjani Prasad, Paidamoyo Chapfuwa +8
Recent advances in immunomics have shown that T-cell receptor (TCR) signatures can accurately predict active or recent infection by leveraging the high specificity of TCR binding t…
Differentially Private Diffusion Models Generate Useful Synthetic Images
Sahra Ghalebikesabi, Leonard Berrada, Sven Gowal +7
The ability to generate privacy-preserving synthetic versions of sensitive image datasets could unlock numerous ML applications currently constrained by data availability. Due to t…
Quasi-Bayesian Nonparametric Density Estimation via Autoregressive Predictive Updates
Sahra Ghalebikesabi, Chris Holmes, Edwin Fong +1
Bayesian methods are a popular choice for statistical inference in small-data regimes due to the regularization effect induced by the prior. In the context of density estimation, t…
Mitigating Statistical Bias within Differentially Private Synthetic Data
Sahra Ghalebikesabi, Harrison Wilde, Jack Jewson +3
Increasing interest in privacy-preserving machine learning has led to new and evolved approaches for generating private synthetic data from undisclosed real data. However, mechanis…
On Locality of Local Explanation Models
Sahra Ghalebikesabi, Lucile Ter-Minassian, Karla Diaz-Ordaz +1
Shapley values provide model agnostic feature attributions for model outcome at a particular instance by simulating feature absence under a global population distribution. The use…
Deep Generative Pattern-Set Mixture Models for Nonignorable Missingness
Sahra Ghalebikesabi, Rob Cornish, Luke J. Kelly +1
We propose a variational autoencoder architecture to model both ignorable and nonignorable missing data using pattern-set mixtures as proposed by Little (1993). Our model explicitl…