mlsanityData to 7 Oct 2026

An Introduction to Machine Unlearning

arXiv 2209.00939 ↗First posted 2 Sep 2022cs.LGcited by 13

On the map: LLM safety and security → Machine unlearning

Abstract (arXiv)

Removing the influence of a specified subset of training data from a machine learning model may be required to address issues such as privacy, fairness, and data quality. Retraining the model from scratch on the remaining data after removal of the subset is an effective but often infeasible option, due to its computational expense. The past few years have therefore seen several novel approaches towards efficient removal, forming the field of "machine unlearning", however, many aspects of the literature published thus far are disparate and lack consensus. In this paper, we summarise and compare seven state-of-the-art machine unlearning algorithms, consolidate definitions of core concepts used in the field, reconcile different approaches for evaluating algorithms, and discuss issues related to applying machine unlearning in practice.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Placed in Machine unlearning, one of 12 directions in LLM safety and security. Builds on 24 of its 60 references that are in the map. First posted on arXiv 2 Sep 2022.

Builds on

24 of 60 references.

Closest ancestors · 6

Same direction, newest first

  1. 2022Knowledge Removal in Sampling-based Bayesian Inference24 Mar
  2. 2022Markov Chain Monte Carlo-Based Machine Unlearning: Unlearning What Needs to be Forgotten28 Feb
  3. 2021On the Necessity of Auditable Algorithmic Definitions for Machine Unlearning22 Oct
  4. 2021Certifiable Machine Unlearning for Linear Models29 Jun
  5. 2021Adaptive Machine Unlearning8 Jun
  6. 2021DeepObliviate: A Powerful Charm for Erasing Data Residual Memory in Deep Neural Networks13 May

Other references · 18

  1. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  2. 2016Matching Networks for One Shot Learningcited by 2,246
  3. 2017Understanding Black-box Predictions via Influence Functionscited by 1,714
  4. 2017Towards A Rigorous Science of Interpretable Machine Learningcited by 1,400
  5. 2019Machine Unlearningcited by 1,020
  6. 2019Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networkscited by 599
  7. 2019Certified Data Removal from Machine Learning Modelscited by 512
  8. 2019Making AI Forget You: Data Deletion in Machine Learningcited by 447
  9. 2020Descent-to-Delete: Gradient-Based Methods for Machine Unlearningcited by 303
  10. 2020Mixed-Privacy Forgetting in Deep Networkscited by 158
  11. 2020DeltaGrad: Rapid retraining of machine learning modelscited by 143
  12. 2020Machine Unlearning for Random Forestscited by 125
6 more
  1. 2020Machine Unlearning: Linear Filtration for Logit-based Classifierscited by 89
  2. 2021Proof-of-Learning: Definitions and Practicecited by 81
  3. 2022Recommendation Unlearningcited by 71
  4. 2022PUMA: Performance Unchanged Model Augmentation for Training Data Removalcited by 56
  5. 2020Towards Probabilistic Verification of Machine Unlearningcited by 46
  6. 2020Coded Machine Unlearningcited by 35

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 6; 2024: 3; 2025: 3.

13 papers.

Sort:

  1. 2022A Survey of Machine Unlearningcited by 298
  2. 2023Machine Unlearning: Solutions and Challengescited by 85
  3. 2023Exploring the Landscape of Machine Unlearning: A Comprehensive Survey and Taxonomycited by 44
  4. 2024Towards Certified Unlearning for Deep Neural Networkscited by 28
  5. 2024Is Retain Set All You Need in Machine Unlearning? Restoring Performance of Unlearned Models with Out-Of-Distribution Imagescited by 18
  6. 2024Unlearning during Learning: An Efficient Federated Machine Unlearning Methodcited by 14
  7. 2023DUCK: Distance-based Unlearning via Centroid Kinematicscited by 13
  8. 2023DeepClean: Machine Unlearning on the Cheap by Resetting Privacy Sensitive Weights using the Fisher Diagonalcited by 7
  9. 2023Machine Unlearning: its nature, scope, and importance for a "delete culture"cited by 5
  10. 2025Position: Bridge the Gaps between Machine Unlearning and AI Regulationcited by 5
  11. 2023Forgettable Federated Linear Learning with Certified Data Unlearningcited by 4
  12. 2025How to Protect Models against Adversarial Unlearning?cited by 0
1 more
  1. 2025Beyond Uniform Deletion: A Data Value-Weighted Framework for Certified Machine Unlearningcited by 0