AI Paper Atlas

Research map / Federated learning and privacy / Privacy attacks and secure learning

Students Parrot Their Teachers: Membership Inference on Model Distillation

Matthew Jagielski, Milad Nasr, Katherine Lee, Christopher A. Choquette-Choo, Nicholas Carlini, Florian Tramèr

NeurIPS 2023 · Federated learning and privacy › Privacy attacks and secure learning · approach: Membership inference and auditing · 6 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
Distill the teacher into a student rather than attack the teacher directlyTeacher training examples · CIFAR-10, WikiText103, Purchase-100, Texas-100Membership inference vulnerability: 5% of examples' vulnerability drop by less than 8 percentage points on CIFAR-10, 5 points on Purchase-100, and 4 points Mixed
Use student queries to infer teacher-example membership without querying on the targetTeacher training examples · CIFAR-10Membership inference accuracy: 73%Improved
Use teacher-label logits instead of student-label logits (ablation)Indirect student-query attack · Teacher-example membership inferenceAttack performanceImproved
Filter student queries to the most informative examples (ablation)Indirect student-query attack · Teacher-example membership inferenceAttack performanceImproved
Use teacher-label logits and filter student queries (ablation)Indirect student-query attack · Teacher-example membership inferenceTPR at FPR: TPR of $10^{-3}$ at a FPR of $10^{-4}$Improved
Use neither teacher-label scoring nor query filtering (ablation)Indirect student-query attack · Teacher-example membership inferenceAttack performanceHurt
Allow near-duplicate examples across teacher and student datasetsIndirect student-query attack · CIFAR-10 teacher examples with student-set duplicatesMembership inference vulnerability: $p<10^{-15}$Hurt
Increase the number of poisoned teacher examplesTeacher-set membership inference · CIFAR-10Membership inference accuracyHurt
Reduce distillation temperature from 4 to 0.1Student model · Teacher-example membership inferenceTPR at FPR: reducing TPR by a factor of roughly 4 at an FPR of $10^{-3}$Improved
Give the attacker less knowledge of the teacher modelPrivate-student membership inference · CIFAR-10Attack success: a TPR as large as $10^{-2}$ at a FPR of $10^{-3}$Improved
Increase the distillation-loss weight alpha in self-distillationSelf-distilled student model · Self-distillationMembership inference attack successHurt
Train a student model from student queries rather than attack the queries aloneTeacher-example membership inference · Teacher training examplesAttack successImproved

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