mlsanityData to 14 Oct 2026

Addressing The Devastating Effects Of Single-Task Data Poisoning In Exemplar-Free Continual Learning

arXiv 2507.04106 ↗First posted 5 Jul 2025cs.CR cs.AIcited by 0

On the map: LLM safety and security → Backdoor and data poisoning attacks

Abstract (arXiv)

Our research addresses the overlooked security concerns related to data poisoning in continual learning (CL). Data poisoning - the intentional manipulation of training data to affect the predictions of machine learning models - was recently shown to be a threat to CL training stability. While existing literature predominantly addresses scenario-dependent attacks, we propose to focus on a more simple and realistic single-task poison (STP) threats. In contrast to previously proposed poisoning settings, in STP adversaries lack knowledge and access to the model, as well as to both previous and future tasks. During an attack, they only have access to the current task within the data stream. Our study demonstrates that even within these stringent conditions, adversaries can compromise model performance using standard image corruptions. We show that STP attacks are able to strongly disrupt the whole continual training process: decreasing both the stability (its performance on past tasks) and plasticity (capacity to adapt to new tasks) of the algorithm. Finally, we propose a high-level defense framework for CL along with a poison task detection method based on task vectors. The code is available at https://github.com/stapaw/STP.git .

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security. Builds on 22 of its 30 references that are in the map. First posted on arXiv 5 Jul 2025.

Builds on

22 of 30 references.

Closest ancestors · 3

Same direction, newest first

  1. 2022Data Poisoning Attack Aiming the Vulnerability of Continual Learning29 Nov
  2. 2022Not All Poisons are Created Equal: Robust Training against Data Poisoning18 Oct
  3. 2022Wild Patterns Reloaded: A Survey of Machine Learning Security against Training Data Poisoning4 May

Other references · 19

  1. 2016Overcoming catastrophic forgetting in neural networkscited by 4,787
  2. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  3. 2016Learning without Forgettingcited by 2,496
  4. 2022Editing Models with Task Arithmeticcited by 1,120
  5. 2019Three scenarios for continual learningcited by 614
  6. 2018Poison Frogs! Targeted Clean-Label Poisoning Attacks on Neural Networkscited by 519
  7. 2020Class-incremental learning: survey and performance evaluation on image classificationcited by 449
  8. 2020Input-Aware Dynamic Backdoor Attackcited by 313
  9. 2018Learning without Memorizingcited by 250
  10. 2017Towards Poisoning of Deep Learning Algorithms with Back-gradient Optimizationcited by 197
  11. 2021SPECTRE: Defending Against Backdoor Attacks Using Robust Statisticscited by 130
  12. 2022FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learningcited by 121
7 more
  1. 2020On the Effectiveness of Mitigating Data Poisoning Attacks with Gradient Shapingcited by 83
  2. 2017Generative Poisoning Attack Method Against Neural Networkscited by 52
  3. 2019On Defending Against Label Flipping Attacks on Malware Detection Systemscited by 22
  4. 2020Targeted Forgetting and False Memory Formation in Continual Learners through Adversarial Backdoor Attackscited by 15
  5. 2018Attack Strength vs. Detectability Dilemma in Adversarial Machine Learningcited by 14
  6. 2023BrainWash: A Poisoning Attack to Forget in Continual Learningcited by 9
  7. 2022False Memory Formation in Continual Learners Through Imperceptible Backdoor Triggercited by 6

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Theory of Continual Learning Against Data Poisoning Attacks
  2. 2023PACOL: Poisoning Attacks Against Continual Learners
  3. 2024Persistent Backdoor Attacks in Continual Learning
  4. 2026Robust Dynamic Expansion for Continual Learning under Backdoor Attacks via Purification and Selective Recovery
  5. 2021Adversarial Targeted Forgetting in Regularization and Generative Based Continual Learning Models
  6. 2023Backdoor Attacks Against Incremental Learners: An Empirical Evaluation Study
  7. 2026Sequential Data Poisoning in LLM Post-Training
  8. 2023Exploring the Limits of Model-Targeted Indiscriminate Data Poisoning Attacks