AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Black-Box Privacy Attacks on Shared Representations in Multitask Learning

John Abascal, Nicolás Berrios, Alina Oprea, Jonathan Ullman, Adam Smith, Matthew Jagielski

ICLR 2026 · Adversarial robustness and security › Black-box and transfer attacks · approach: Decision-based boundary attacks · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
give adversary the actual training samples (strong) instead of fresh samples (weak)inner product attack · CelebA personalization, 50th percentilebalanced accuracy: 52.4% -> 80%Improved
strong instead of weak adversaryinner product attack · Stack Overflow personalization, 50th percentilebalanced accuracy: 58.2% -> 98.7%Improved
strong instead of weak adversaryvariance attack · CelebA personalization, FPR 1%TPR: 2.9% -> 61.2%Improved
strong instead of weak adversaryinner product attack · Stack Overflow personalization, FPR 1%TPR: 8.2% -> 98.5%Improved
define tasks as labelled topics (multiple learning problems) instead of usersweak adversary inner product attack · Stack Overflow, 50th percentilebalanced accuracy: 58.2% -> 85.9%Improved
define tasks as users (personalization) rather than labelled topicsstrong vs weak gap · Stack Overflowbalanced accuracy: 98.7% -> 87%Mixed
raise percentile threshold from 50th to 90thinner product attack strong adversary · CelebA personalizationbalanced accuracy: 80% -> 59.5%Hurt
increase total samples (samples per task or number of tasks) (ablation)variance attack strong adversary · synthetic dataAUCImproved
increase embedding dimension (ablation)variance attack · synthetic dataAUCNo clear effect
larger generalization gap (train loss minus zero-shot loss)strong adversary inner product attack · Stack OverflowAUCImproved
use gradient clipping and normalization as defenseMTL training · task-inference attacksattack successNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy