Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Continual Learning through Networks Splitting and Merging with Dreaming-Meta-Weighted Model Fusion (2312.07082v1)

Published 12 Dec 2023 in cs.CV

Abstract: It's challenging to balance the networks stability and plasticity in continual learning scenarios, considering stability suffers from the update of model and plasticity benefits from it. Existing works usually focus more on the stability and restrict the learning plasticity of later tasks to avoid catastrophic forgetting of learned knowledge. Differently, we propose a continual learning method named Split2MetaFusion which can achieve better trade-off by employing a two-stage strategy: splitting and meta-weighted fusion. In this strategy, a slow model with better stability, and a fast model with better plasticity are learned sequentially at the splitting stage. Then stability and plasticity are both kept by fusing the two models in an adaptive manner. Towards this end, we design an optimizer named Task-Preferred Null Space Projector(TPNSP) to the slow learning process for narrowing the fusion gap. To achieve better model fusion, we further design a Dreaming-Meta-Weighted fusion policy for better maintaining the old and new knowledge simultaneously, which doesn't require to use the previous datasets. Experimental results and analysis reported in this work demonstrate the superiority of the proposed method for maintaining networks stability and keeping its plasticity. Our code will be released.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (56)
  1. Memory aware synapses: Learning what (not) to forget. In Proceedings of the European conference on computer vision (ECCV), pages 139–154, 2018.
  2. Coresets via bilevel optimization for continual learning and streaming. Advances in Neural Information Processing Systems, 33:14879–14890, 2020.
  3. Modeling the background for incremental learning in semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9233–9242, 2020.
  4. Efficient lifelong learning with a-gem. arXiv preprint arXiv:1812.00420, 2018.
  5. Using hindsight to anchor past knowledge in continual learning. In Proceedings of the AAAI conference on artificial intelligence, pages 6993–7001, 2021.
  6. Mitigating forgetting in online continual learning via instance-aware parameterization. Advances in Neural Information Processing Systems, 33:17466–17477, 2020.
  7. Continual prototype evolution: Learning online from non-stationary data streams. In Proceedings of the IEEE/CVF international conference on computer vision, pages 8250–8259, 2021.
  8. Flattening sharpness for dynamic gradient projection memory benefits continual learning. Advances in Neural Information Processing Systems, 34:18710–18721, 2021.
  9. Robust mean teacher for continual and gradual test-time adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7704–7714, 2023.
  10. Plop: Learning without forgetting for continual semantic segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4040–4050, 2021.
  11. Adversarial continual learning. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XI 16, pages 386–402. Springer, 2020.
  12. The pascal visual object classes challenge: A retrospective. International journal of computer vision, 111:98–136, 2015.
  13. Cosda: Continual source-free domain adaptation. arXiv preprint arXiv:2304.06627, 2023.
  14. Qualitatively characterizing neural network optimization problems. arXiv preprint arXiv:1412.6544, 2014.
  15. Online continual learning through mutual information maximization. In International Conference on Machine Learning, pages 8109–8126. PMLR, 2022.
  16. Natural continual learning: success is a journey, not (just) a destination. Advances in neural information processing systems, 34:28067–28079, 2021.
  17. Achieving forgetting prevention and knowledge transfer in continual learning. Advances in Neural Information Processing Systems, 34:22443–22456, 2021.
  18. Overcoming catastrophic forgetting in neural networks. Proceedings of the national academy of sciences, 114(13):3521–3526, 2017.
  19. Balancing stability and plasticity through advanced null space in continual learning. In European Conference on Computer Vision, pages 219–236. Springer, 2022.
  20. Learning multiple layers of features from tiny images. 2009.
  21. Overcoming catastrophic forgetting with unlabeled data in the wild. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 312–321, 2019.
  22. Visualizing the loss landscape of neural nets. Advances in neural information processing systems, 31, 2018.
  23. Interpretable neural network decoupling. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XV 16, pages 653–669. Springer, 2020.
  24. Learning without forgetting. IEEE transactions on pattern analysis and machine intelligence, 40(12):2935–2947, 2017.
  25. Adaptive plasticity improvement for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7816–7825, 2023.
  26. Towards better plasticity-stability trade-off in incremental learning: A simple linear connector. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 89–98, 2022.
  27. Gradient episodic memory for continual learning. Advances in neural information processing systems, 30, 2017.
  28. Recall: Replay-based continual learning in semantic segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 7026–7035, 2021.
  29. Alleviating catastrophic forgetting using context-dependent gating and synaptic stabilization. Proceedings of the National Academy of Sciences, 115(44):E10467–E10475, 2018.
  30. Catastrophic interference in connectionist networks: The sequential learning problem. In Psychology of learning and motivation, pages 109–165. Elsevier, 1989.
  31. Incremental learning techniques for semantic segmentation. In Proceedings of the IEEE/CVF international conference on computer vision workshops, pages 0–0, 2019.
  32. Linear mode connectivity in multitask and continual learning. arXiv preprint arXiv:2010.04495, 2020.
  33. D3former: Debiased dual distilled transformer for incremental learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2420–2429, 2023.
  34. Inceptionism: Going deeper into neural networks. 2015.
  35. On first-order meta-learning algorithms. arXiv preprint arXiv:1803.02999, 2018.
  36. Overcoming catastrophic forgetting by neuron-level plasticity control. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 5339–5346, 2020.
  37. icarl: Incremental classifier and representation learning. In Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, pages 2001–2010, 2017.
  38. Learning to learn without forgetting by maximizing transfer and minimizing interference. In International Conference on Learning Representations.
  39. Gradient projection memory for continual learning. In International Conference on Learning Representations, 2020.
  40. Zipit! merging models from different tasks without training. arXiv preprint arXiv:2305.03053, 2023.
  41. Disparse: Disentangled sparsification for multitask model compression. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12382–12392, 2022a.
  42. Meta-gf: Training dynamic-depth neural networks harmoniously. In European Conference on Computer Vision, pages 691–708. Springer, 2022b.
  43. Learning task-preferred inference routes for gradient de-conflict in multi-output dnns. arXiv preprint arXiv:2305.19844, 2023.
  44. Topology-preserving class-incremental learning. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XIX 16, pages 254–270. Springer, 2020.
  45. Hippocampal offline reactivation consolidates recently formed cell assembly patterns during sharp wave-ripples. Neuron, 92(5):968–974, 2016.
  46. Brain-inspired replay for continual learning with artificial neural networks. Nature communications, 11(1):4069, 2020.
  47. Three types of incremental learning. Nature Machine Intelligence, 4(12):1185–1197, 2022.
  48. Rehearsal revealed: The limits and merits of revisiting samples in continual learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9385–9394, 2021.
  49. Afec: Active forgetting of negative transfer in continual learning. Advances in Neural Information Processing Systems, 34:22379–22391, 2021a.
  50. Training networks in null space of feature covariance for continual learning. In Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition, pages 184–193, 2021b.
  51. Online continual learning with contrastive vision transformer. In Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XX, pages 631–650. Springer, 2022a.
  52. Dualprompt: Complementary prompting for rehearsal-free continual learning. In European Conference on Computer Vision, pages 631–648. Springer, 2022b.
  53. Tiny imagenet challenge. Technical report, 2017.
  54. Dreaming to distill: Data-free knowledge transfer via deepinversion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8715–8724, 2020.
  55. Gradient norm aware minimization seeks first-order flatness and improves generalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20247–20257, 2023.
  56. Relationship between dreaming and memory reconsolidation. Brain Science Advances, 4(2):118–130, 2018.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (6)
  1. Yi Sun (146 papers)
  2. Xin Xu (187 papers)
  3. Jian Li (667 papers)
  4. Guanglei Xie (1 paper)
  5. Yifei Shi (26 papers)
  6. Qiang Fang (16 papers)