mlsanityData to 7 Oct 2026

Formalizing the Generalization-Forgetting Trade-off in Continual Learning

arXiv 2109.14035 ↗First posted 28 Sep 2021cs.LG cs.AI math.DScited by 24

Not on the map. Why?

Abstract (arXiv)

We formulate the continual learning (CL) problem via dynamic programming and model the trade-off between catastrophic forgetting and generalization as a two-player sequential game. In this approach, player 1 maximizes the cost due to lack of generalization whereas player 2 minimizes the cost due to catastrophic forgetting. We show theoretically that a balance point between the two players exists for each task and that this point is stable (once the balance is achieved, the two players stay at the balance point). Next, we introduce balanced continual learning (BCL), which is designed to attain balance between generalization and forgetting and empirically demonstrate that BCL is comparable to or better than the state of the art.

Key facts. Cited by 24 papers in the map (data through 7 Oct 2026). Builds on 36 of its 63 references that are in the map. First posted on arXiv 28 Sep 2021.

Builds on

36 of 63 references.

References · 36

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  3. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  4. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  5. 2016Learning without Forgettingcited by 2,465
  6. 2016Matching Networks for One Shot Learningcited by 2,246
  7. 2017Gradient Episodic Memory for Continual Learningcited by 1,871
  8. 2016Progressive Neural Networkscited by 1,332
  9. 2017Continual Learning with Deep Generative Replaycited by 1,152
  10. 2017Memory Aware Synapses: Learning what (not) to forgetcited by 1,094
  11. 2018On First-Order Meta-Learning Algorithmscited by 1,003
  12. 2017Lifelong Learning with Dynamically Expandable Networkscited by 689
24 more
  1. 2018Progress & Compress: A scalable framework for continual learningcited by 516
  2. 2018Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interferencecited by 511
  3. 2019Orthogonal Gradient Descent for Continual Learningcited by 361
  4. 2019On Tiny Episodic Memories in Continual Learningcited by 299
  5. 2018Meta-Learning: A Surveycited by 254
  6. 2018Re-evaluating Continual Learning Scenarios: A Categorization and Case for Strong Baselinescited by 228
  7. 2019Meta-Learning Representations for Continual Learningcited by 189
  8. 2020Understanding the Role of Training Regimes in Continual Learningcited by 177
  9. 2019Functional Regularisation for Continual Learning with Gaussian Processescited by 121
  10. 2020Continual Deep Learning by Functional Regularisation of Memorable Pastcited by 104
  11. 2018Generative replay with feedback connections as a general strategy for continual learningcited by 103
  12. 2020Continual Learning with Node-Importance based Adaptive Group Sparse Regularizationcited by 102
  13. 2020A Theoretical Analysis of Catastrophic Forgetting through the NTK Overlap Matrixcited by 99
  14. 2020Continual Learning in Low-rank Orthogonal Subspacescited by 99
  15. 2020Optimal Continual Learning has Perfect Memory and is NP-hardcited by 86
  16. 2020Learning to Continually Learncited by 82
  17. 2018Deep Online Learning via Meta-Learning: Continual Adaptation for Model-Based RLcited by 80
  18. 2020Generalisation Guarantees for Continual Learning with Orthogonal Gradient Descentcited by 68
  19. 2020La-MAML: Look-ahead Meta Learning for Continual Learningcited by 51
  20. 2020Online Continual Learning under Extreme Memory Constraintscited by 48
  21. 2020Meta-Consolidation for Continual Learningcited by 40
  22. 2020Online Fast Adaptation and Knowledge Accumulation: a New Approach to Continual Learningcited by 38
  23. 2018Meta Continual Learningcited by 4
  24. 2021CoachNet: An Adversarial Sampling Approach for Reinforcement Learningcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 7; 2023: 6; 2024: 1; 2025: 6; 2026: 3.

24 papers.

Sort:

By area:

  1. 2023On the Stability-Plasticity Dilemma of Class-Incremental Learningcited by 55
  2. 2023Active Neural Mappingcited by 30
  3. 2022How to Reuse and Compose Knowledge for a Lifetime of Tasks: A Survey on Continual Learning and Functional Compositioncited by 27
  4. 2022Task-Free Continual Learning via Online Discrepancy Distance Learningcited by 22
  5. 2022Continual Variational Autoencoder Learning via Online Cooperative Memorizationcited by 12
  6. 2025Plasticity as the Mirror of Empowermentcited by 11
  7. 2022FedKNOW: Federated Continual Learning with Signature Task Knowledge Integration at Edgecited by 10
  8. 2022Hierarchically Structured Task-Agnostic Continual Learningcited by 9
  9. 2022Theoretical Understanding of the Information Flow on Continual Learning Performancecited by 8
  10. 2021Mixture-of-Variational-Experts for Continual Learningcited by 7
  11. 2023IBCL: Zero-shot Model Generation under Stability-Plasticity Trade-offscited by 5
  12. 2023Doubly Perturbed Task Free Continual Learningcited by 5
12 more
  1. 2025Sequence Transferability and Task Order Selection in Continual Learningcited by 5
  2. 2024Parameter-Efficient Fine-Tuning for Continual Learning: A Neural Tangent Kernel Perspectivecited by 2
  3. 2022Increasing Depth of Neural Networks for Life-long Learningcited by 1
  4. 2023Analysis of the Memorization and Generalization Capabilities of AI Agents: Are Continual Learners Robust?cited by 1
  5. 2025On Understanding of the Dynamics of Model Capacity in Continual Learningcited by 1
  6. 2026Mitigating Task-Order Sensitivity and Forgetting via Hierarchical Second-Order Consolidationcited by 1
  7. 2026Cross-Sample Relational Fusion: Unifying Domain Generalization and Class-Incremental Learningcited by 1
  8. 2023Learning Continually on a Sequence of Graphs -- The Dynamical System Waycited by 0
  9. 2025Ergodic Risk Measures: Towards a Risk-Aware Foundation for Continual Reinforcement Learningcited by 0
  10. 2025Forgetting is Everywherecited by 0
  11. 2025Erase to Retain: Low Rank Adaptation Guided Selective Unlearning in Medical Segmentation Networkscited by 0
  12. 2026Understanding the Role of Rehearsal Scale in Continual Learning under Varying Model Capacitiescited by 0