Data to 7 Oct 2026

Radioactive data: tracing through training

arXiv 2002.00937 ↗First posted 3 Feb 2020stat.ML cs.CR cs.CVcited by 67

Not on the map. Why?

Cited by 25 in LLM safety and security, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

We want to detect whether a particular image dataset has been used to train a model. We propose a new technique, radioactive data, that makes imperceptible changes to this dataset such that any model trained on it will bear an identifiable mark. The mark is robust to strong variations such as different architectures or optimization methods. Given a trained model, our technique detects the use of radioactive data and provides a level of confidence (p-value). Our experiments on large-scale benchmarks (Imagenet), using standard architectures (Resnet-18, VGG-16, Densenet-121) and training procedures, show that we can detect usage of radioactive data with high confidence (p<10^-4) even when only 1% of the data used to trained our model is radioactive. Our method is robust to data augmentation and the stochasticity of deep network optimization. As a result, it offers a much higher signal-to-noise ratio than data poisoning and backdoor methods.

Key facts. Cited by 67 papers in the map (data through 7 Oct 2026). First posted on arXiv 3 Feb 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 6; 2022: 8; 2023: 8; 2024: 17; 2025: 16; 2026: 12.

67 papers.

Sort:

By area:

2024Moshi: a speech-text foundation model for real-time dialoguecited by 536
2022Untargeted Backdoor Watermark: Towards Harmless and Stealthy Dataset Copyright Protectioncited by 80
2024Watermarking Makes Language Models Radioactivecited by 52
2021CoProtector: Protect Open-Source Code against Unauthorized Training Usage with Data Poisoningcited by 39
2024Proving membership in LLM pretraining data via data watermarkscited by 38
2023Did You Train on My Dataset? Towards Public Dataset Protection with Clean-Label Backdoor Watermarkingcited by 37
2022Understanding Rare Spurious Correlations in Neural Networkscited by 27
2024Towards Reliable Verification of Unauthorized Data Usage in Personalized Text-to-Image Diffusion Modelscited by 19
2021A Framework for Deprecating Datasets: Standardizing Documentation, Identification, and Communicationcited by 18
2021SoK: Anti-Facial Recognition Technologycited by 16
2023GrOVe: Ownership Verification of Graph Neural Networks using Embeddingscited by 16
2024ProMark: Proactive Diffusion Watermarking for Causal Attributioncited by 16

38 more

2024A General Framework for Data-Use Auditing of ML Modelscited by 15
2024Ward: Provable RAG Dataset Inference via LLM Watermarkscited by 15
2024SoK: Dataset Copyright Auditing in Machine Learning Systemscited by 14
2022Conflicting Interactions Among Protection Mechanisms for Machine Learning Modelscited by 12
2024Is My Data in Your AI? Membership Inference Test (MINT) applied to Face Biometricscited by 12
2022Holistic Adversarial Robustness of Deep Learning Modelscited by 11
2022On the Effectiveness of Dataset Watermarking in Adversarial Settingscited by 11
2022Data Isotopes for Data Provenance in DNNscited by 11
2023Identifying Appropriate Intellectual Property Protection Mechanisms for Machine Learning Models: A Systematization of Watermarking, Fingerprinting, Model Access, and Attackscited by 11
2025Unified attacks to large language model watermarks: spoofing and scrubbing in unauthorized knowledge distillationcited by 9
2022On the Robustness of Dataset Inferencecited by 8
2023GraphGuard: Detecting and Counteracting Training Data Misuse in Graph Neural Networkscited by 8
2024Data Taggants: Dataset Ownership Verification via Harmless Targeted Data Poisoningcited by 8
2025DuFFin: A Dual-Level Fingerprinting Framework for LLMs IP Protectioncited by 8
2025DATABench: Evaluating Dataset Auditing in Deep Learning from an Adversarial Perspectivecited by 8
2023SoK: Unintended Interactions among Machine Learning Defenses and Riskscited by 6
2025Detecting Benchmark Contamination Through Watermarkingcited by 6
2025Winter Soldier: Backdooring Language Models at Pre-Training with Indirect Data Poisoningcited by 6
2025Radioactive Watermarks in Diffusion and Autoregressive Image Generative Modelscited by 6
202110 Security and Privacy Problems in Large Foundation Modelscited by 5
2023Adversarial Machine Learning for Social Good: Reframing the Adversary as an Allycited by 5
2025BitMark: Watermarking Bitwise Autoregressive Image Generative Modelscited by 5
2021Anti-Neuron Watermarking: Protecting Personal Data Against Unauthorized Neural Networkscited by 4
2021Lightweight machine unlearning in neural networkcited by 4
2024DeepEclipse: How to Break White-Box DNN-Watermarking Schemescited by 4
2025Obliviate: Efficient Unmemorization for Protecting Intellectual Property in Large Language Modelscited by 4
2025Instance-Level Data-Use Auditing of Visual ML Modelscited by 4
2023Hey That's Mine Imperceptible Watermarks are Preserved in Diffusion Generated Outputscited by 3
2024MemControl: Mitigating Memorization in Diffusion Models via Automated Parameter Selectioncited by 3
2024Towards Data Governance of Frontier AI Modelscited by 3
2025Attribution-by-design: Ensuring Inference-Time Provenance in Generative Music Systemscited by 3
2026Watermarking LLM Agent Trajectoriescited by 3
2024Proactive Schemes: A Survey of Adversarial Attacks for Social Goodcited by 2
2024Combining Machine Learning Defenses without Conflictscited by 2
2025A Survey on Foundation Models for Personalized Federated Intelligencecited by 2
2025Evading Data Provenance in Deep Neural Networkscited by 2
2025SSCL-BW: Sample-Specific Clean-Label Backdoor Watermarking for Dataset Ownership Verificationcited by 2
2026TextSeal: A Localized LLM Watermark for Provenance & Distillation Protectioncited by 2

50 most cited of 67 All 67 (JSON)