Privacy Attacks papers, 2023–2026

Key findings

• Privacy Attacks went from 124 papers in 2023 to 219 in 2026, roughly 1.8x in three years.
• ICLR publishes the most Privacy Attacks work: 262 papers from 2023 to 2026, 38% of the four-venue total.
• The fastest-growing overlap is with Alignment and Safety: papers tagged with both rose from 2 in 2023 to 80 in 2026.
• Among the 5 topics under Federated and Privacy, it ranks #1 by 2026 paper count; the fastest grower in the group, Secure Aggregation, is up 1.8x versus 1.8x here.

Representative papers

Spread across years and venues so every edition shows up

• NeurIPS 2026A Private Empirical Defense Against Privacy Audits
• ICML 2026A Statistical Framework for Analyzing Specification Resistance to Learnware-Inversion Risks
• ICLR 2026(Token-Level) \textbf{InfoRMIA}: Stronger Membership Inference and Privacy Assessment for LLMs
• CVPR 2026Out of Sight Out of Track Adversarial Attacks on Propagation-based
• NeurIPS 2025A Reliable Cryptographic Framework for Empirical Machine Unlearning Evaluation
• ICML 2025A Certified Unlearning Approach without Access to Source Data
• ICLR 2025Adversarial Attacks on Data Attribution
• CVPR 2025Adv-CPG: A Customized Portrait Generation Framework with Facial Adversarial Attacks
• NeurIPS 2024Reproducibility Study on Adversarial Attacks Against Robust Transformer Trackers
• ICML 2024Collective Certified Robustness against Graph Injection Attacks
• ICLR 2024Be Careful What You Smooth For: Label Smoothing Can Be a Privacy Shield but Also a Catalyst for Model Inversion Attacks
• CVPR 2024Efficient Privacy-Preserving Visual Localization Using 3D Ray Clouds
• NeurIPS 2023Bounding the Invertibility of Privacy-preserving Instance Encoding using Fisher Information
• ICML 2023Data Poisoning Attacks Against Multimodal Encoders
• ICLR 2023Adversarial Attacks on Adversarial Bandits
• CVPR 2023Can't Steal? Cont-Steal! Contrastive Stealing Attacks Against Image Encoders
• NeurIPS 2026Adaptive Target-Charging with Privacy Filters and Individual Accounting
• ICML 2026How does Bayesian Sampling help Membership Inference Attacks?
• ICLR 2026Beyond Membership: Limitations of Add/Remove Adjacency in Differential Privacy
• CVPR 2026Revisiting Geometric Obfuscation with Dual Convergent Lines for Privacy-Preserving Image

All Privacy Attacks papers

698 papers, page 7 of 14

301. ICLR 2025Scalable Extraction of Training Data from Aligned, Production Language Models
302. ICLR 2025SelectFormer in Data Markets: Privacy-Preserving and Efficient Data Selection for Transformers with Multi-Party Computation
303. ICLR 2025SimPER: A Minimalist Approach to Preference Alignment without Hyperparameters
304. ICLR 2025Singular Subspace Perturbation Bounds via Rectangular Random Matrix Diffusions
305. ICLR 2025Stealthy Shield Defense: A Conditional Mutual Information-Based Approach against Black-Box Model Inversion Attacks
306. ICLR 2025Stochastic Bandits Robust to Adversarial Attacks
307. ICLR 2025Support is All You Need for Certified VAE Training
308. ICLR 2025Tamper-Resistant Safeguards for Open-Weight LLMs
309. ICLR 2025Targeted Attack Improves Protection against Unauthorized Diffusion Customization
310. ICLR 2025The Last Iterate Advantage: Empirical Auditing and Principled Heuristic Analysis of Differentially Private SGD
311. ICLR 2025Tighter Privacy Auditing of DP-SGD in the Hidden State Threat Model
312. ICLR 2025To Tackle Adversarial Transferability: A Novel Ensemble Training Method with Fourier Transformation
313. ICLR 2025Towards Certification of Uncertainty Calibration under Adversarial Attacks
314. ICLR 2025Tracking the Copyright of Large Vision-Language Models through Parameter Learning Adversarial Images
315. ICLR 2025Training Robust Ensembles Requires Rethinking Lipschitz Continuity
316. ICLR 2025Understanding and Enhancing the Transferability of Jailbreaking Attacks
317. ICLR 2025Unlearn and Burn: Adversarial Machine Unlearning Requests Destroy Model Accuracy
318. ICLR 2025Unlearning or Obfuscating? Jogging the Memory of Unlearned LLMs via Benign Relearning
319. ICLR 2025Variational Best-of-N Alignment
320. ICLR 2025Wicked Oddities: Selectively Poisoning for Effective Clean-Label Backdoor Attacks
321. ICML 2025A Certified Unlearning Approach without Access to Source Data
322. ICML 2025A Closer Look at Backdoor Attacks on CLIP
323. ICML 2025Auditing Prompt Caching in Language Model APIs
324. ICML 2025Black-Box Adversarial Attacks on LLM-Based Code Completion
325. ICML 2025Connecting Thompson Sampling and UCB: Towards More Efficient Trade-offs Between Privacy and Regret
326. ICML 2025De-AntiFake: Rethinking the Protective Perturbations Against Voice Cloning Attacks
327. ICML 2025Differential Privacy Guarantees of Markov Chain Monte Carlo Algorithms
328. ICML 2025EncryptedLLM: Privacy-Preserving Large Language Model Inference via GPU-Accelerated Fully Homomorphic Encryption
329. ICML 2025Flexible, Efficient, and Stable Adversarial Attacks on Machine Unlearning
330. ICML 2025Fragments to Facts: Partial-Information Fragment Inference from LLMs
331. ICML 2025HyperNear: Unnoticeable Node Injection Attacks on Hypergraph Neural Networks
332. ICML 2025ICLShield: Exploring and Mitigating In-Context Learning Backdoor Attacks
333. ICML 2025Improving the Variance of Differentially Private Randomized Experiments through Clustering
334. ICML 2025Interpreting the Repeated Token Phenomenon in Large Language Models
335. ICML 2025Language Models May Verbatim Complete Text They Were Not Explicitly Trained On
336. ICML 2025Leveraging Per-Instance Privacy for Machine Unlearning
337. ICML 2025MELON: Provable Defense Against Indirect Prompt Injection Attacks in AI Agents
338. ICML 2025On Differential Privacy for Adaptively Solving Search Problems via Sketching
339. ICML 2025Position: Contextual Integrity is Inadequately Applied to Language Models
340. ICML 2025Privacy Attacks on Image AutoRegressive Models
341. ICML 2025Private Lossless Multiple Release
342. ICML 2025Revealing Weaknesses in Text Watermarking Through Self-Information Rewrite Attacks
343. ICML 2025Textual Unlearning Gives a False Sense of Unlearning
344. ICML 2025The Canary’s Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
345. ICML 2025The Ripple Effect: On Unforeseen Complications of Backdoor Attacks
346. ICML 2025The Role of Randomness in Stability
347. ICML 2025Theoretically Unmasking Inference Attacks Against LDP-Protected Clients in Federated Vision Models
348. ICML 2025Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond
349. ICML 2025Underestimated Privacy Risks for Minority Populations in Large Language Model Unlearning
350. NeurIPS 2025A Reliable Cryptographic Framework for Empirical Machine Unlearning Evaluation

FAQ

How many Privacy Attacks papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

698 in total: 124 in 2023, 180 in 2024, 175 in 2025, 219 in 2026.

How many Privacy Attacks papers are at NeurIPS 2026?

NeurIPS 2026 has 55 Privacy Attacks papers, versus 45 the year before.

How many Privacy Attacks papers are at ICLR 2026?

ICLR 2026 has 102 Privacy Attacks papers, versus 83 the year before.

How many Privacy Attacks papers are at ICML 2026?

ICML 2026 has 38 Privacy Attacks papers, versus 29 the year before.

How many Privacy Attacks papers are at CVPR 2026?

CVPR 2026 has 24 Privacy Attacks papers, versus 18 the year before.

Which conference publishes the most Privacy Attacks papers?

ICLR, with 262 papers from 2023 to 2026.