Research map / Adversarial robustness and security / Attacks on text and graphs
Rethinking Membership Inference Attacks for CLIP
Lluis Gomez
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use same-distribution (CommonPool) non-members instead of cross-dataset non-members | CSA, AEA, WSA, MCDropout attacks · CLIP-MIA-Bench vs cross-dataset | AUC: MCDropout 98.27 -> 52.63; WSA 93.49 -> 50.07 | Hurt |
| Increase CLIP training set from 4M to 400M pairs (ablation) | Cosine Similarity Attack · CLIP-MIA-Bench | AUC: 94.1 -> 62.5 -> 51.1 (%) | Hurt |
| Use white-box loss+grad-norm features instead of similarity | MIA on CLIP 400M · CLIP-MIA-Bench | AUC: 51.12 -> 53.48 | Improved |
Closest papers by meaning
- Membership Inference Attacks against Large Vision-Language Models (NeurIPS 2024 · similarity 0.91)
- Causal Evaluation of Membership Inference Attacks (NeurIPS 2026 · similarity 0.91)
- Membership Inference Attacks against Fine-tuned Large Language Models via Self-prompt Calibration (NeurIPS 2024 · similarity 0.90)
- Exploring the limits of strong membership inference attacks on large language models (NeurIPS 2025 · similarity 0.89)
- MIA-Tuner: Adapting Large Language Models as Pre-training Text Detector (AAAI 2025 · similarity 0.89)
- Assessing Per-Sample Membership Inference Vulnerability without Retraining (NeurIPS 2026 · similarity 0.88)
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.
