mlsanityData to 14 Oct 2026

Tight Bounds for Machine Unlearning via Differential Privacy

arXiv 2309.00886 ↗First posted 2 Sep 2023cs.LGcited by 18

On the map: LLM safety and security → Machine unlearning

Abstract (arXiv)

We consider the formulation of "machine unlearning" of Sekhari, Acharya, Kamath, and Suresh (NeurIPS 2021), which formalizes the so-called "right to be forgotten" by requiring that a trained model, upon request, should be able to "unlearn" a number of points from the training data, as if they had never been included in the first place. Sekhari et al. established some positive and negative results about the number of data points that can be successfully unlearnt by a trained model without impacting the model's accuracy (the "deletion capacity"), showing that machine unlearning could be achieved by using differentially private (DP) algorithms. However, their results left open a gap between upper and lower bounds on the deletion capacity of these algorithms: our work fully closes this gap, obtaining tight bounds on the deletion capacity achievable by DP-based machine unlearning algorithms.

Key facts. Cited by 18 papers in the map (data through 14 Oct 2026). Placed in Machine unlearning, one of 12 directions in LLM safety and security. Builds on 16 of its 22 references that are in the map. First posted on arXiv 2 Sep 2023.

Builds on

16 of 22 references.

Closest ancestors · 5

Same direction, newest first

  1. 2022Algorithms that Approximate Data Removal: New Results and Limitations25 Sep
  2. 2022A Survey of Machine Unlearning6 Sep
  3. 2022Deletion Inference, Reconstruction, and Compliance in Machine (Un)Learning7 Feb
  4. 2021Remember What You Want to Forget: Algorithms for Machine Unlearning4 Mar
  5. 2021Machine Unlearning via Algorithmic Stability25 Feb

Other references · 11

  1. 2019Machine Unlearningcited by 1,061
  2. 2019Certified Data Removal from Machine Learning Modelscited by 535
  3. 2019Making AI Forget You: Data Deletion in Machine Learningcited by 465
  4. 2016Concentrated Differential Privacy: Simplifications, Extensions, and Lower Boundscited by 346
  5. 2020Approximate Data Deletion from Machine Learning Modelscited by 255
  6. 2020When Machine Unlearning Jeopardizes Privacycited by 198
  7. 2019Private Stochastic Convex Optimization with Optimal Ratescited by 186
  8. 2020Private Stochastic Convex Optimization: Optimal Rates in Linear Timecited by 153
  9. 2019Updates-Leak: Data Set Inference and Reconstruction Attacks in Online Learningcited by 122
  10. 2015Between Pure and Approximate Differential Privacycited by 56
  11. 2020Formalizing Data Deletion in the Context of the Right to be Forgottencited by 50

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 6; 2025: 6; 2026: 5.

18 papers.

Sort:

By area:

  1. 2023In-Context Unlearning: Language Models as Few Shot Unlearnerscited by 224
  2. 2024Machine Unlearning of Pre-trained Large Language Modelscited by 119
  3. 2024Position: LLM Unlearning Benchmarks are Weak Measures of Progresscited by 63
  4. 2024Machine Unlearning Fails to Remove Data Poisoning Attackscited by 37
  5. 2024The Utility and Complexity of in- and out-of-Distribution Machine Unlearningcited by 20
  6. 2024Gone but Not Forgotten: Improved Benchmarks for Machine Unlearningcited by 9
  7. 2025When to Forget? Complexity Trade-offs in Machine Unlearningcited by 9
  8. 2024Certification for Differentially Private Prediction in Gradient-Based Trainingcited by 7
  9. 2025Position: Bridge the Gaps between Machine Unlearning and AI Regulationcited by 7
  10. 2025SEPS: A Separability Measure for Robust Unlearning in LLMscited by 5
  11. 2025Exploring Nonlinear Pathway in Parameter Space for Machine Unlearningcited by 4
  12. 2025Fully Decentralized Certified Unlearningcited by 1
6 more
  1. 2026Less Noise, Same Certificate: Retain Sensitivity for Unlearningcited by 1
  2. 2025Hierarchical Dual-Strategy Unlearning for Biomedical and Healthcare Intelligence Using Imperfect and Privacy-Sensitive Medical Datacited by 0
  3. 2026Variance-Reduced (varepsilon,δ)-Unlearning using Forget Set Gradientscited by 0
  4. 2026Near-Optimal Machine Unlearning Utility for Smooth Strongly Convex Lossescited by 0
  5. 2026Memorization Is Not Extraction: Tight Differential-Privacy Bounds and Audit Blind Spotscited by 0
  6. 2026Minimax Gaussian Mechanisms for Continual Machine Unlearningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Efficient Unlearning with Privacy Guarantees
  2. 2025Auditing Approximate Machine Unlearning for Differentially Private Models
  3. 2022Forget Unlearning: Towards True Data-Deletion in Machine Learning
  4. 2024Rewind-to-Delete: Certified Machine Unlearning for Nonconvex Functions
  5. 2023Learn to Unlearn: A Survey on Machine Unlearning
  6. 2026Protecting the Undeleted in Machine Unlearning
  7. 2026Mitigating Privacy Risk via Forget Set-Free Unlearning
  8. 2023Fair Machine Unlearning: Data Removal while Mitigating Disparities