Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
80 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 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

Consistent Prompting for Rehearsal-Free Continual Learning (2403.08568v2)

Published 13 Mar 2024 in cs.CV and cs.LG

Abstract: Continual learning empowers models to adapt autonomously to the ever-changing environment or data streams without forgetting old knowledge. Prompt-based approaches are built on frozen pre-trained models to learn the task-specific prompts and classifiers efficiently. Existing prompt-based methods are inconsistent between training and testing, limiting their effectiveness. Two types of inconsistency are revealed. Test predictions are made from all classifiers while training only focuses on the current task classifier without holistic alignment, leading to Classifier inconsistency. Prompt inconsistency indicates that the prompt selected during testing may not correspond to the one associated with this task during training. In this paper, we propose a novel prompt-based method, Consistent Prompting (CPrompt), for more aligned training and testing. Specifically, all existing classifiers are exposed to prompt training, resulting in classifier consistency learning. In addition, prompt consistency learning is proposed to enhance prediction robustness and boost prompt selection accuracy. Our Consistent Prompting surpasses its prompt-based counterparts and achieves state-of-the-art performance on multiple continual learning benchmarks. Detailed analysis shows that improvements come from more consistent training and testing.

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 (51)
  1. Memory aware synapses: Learning what (not) to forget. In Proceedings of the European conference on computer vision (ECCV), pages 139–154, 2018.
  2. A comprehensive study of class incremental learning algorithms for visual tasks. Neural Networks, 135:38–54, 2021.
  3. On the effectiveness of lipschitz-driven rehearsal in continual learning. Advances in Neural Information Processing Systems, 35:31886–31901, 2022.
  4. Dark experience for general continual learning: a strong, simple baseline. Advances in neural information processing systems, 33:15920–15930, 2020.
  5. Dynamic residual classifier for class incremental learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 18743–18752, 2023.
  6. A continual learning survey: Defying forgetting in classification tasks. IEEE transactions on pattern analysis and machine intelligence, 44(7):3366–3385, 2021.
  7. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020.
  8. Overcoming catastrophic forgetting in incremental object detection via elastic response distillation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9427–9436, 2022.
  9. R-dfcil: Relation-guided representation learning for data-free class incremental learning. In European Conference on Computer Vision, pages 423–439. Springer, 2022.
  10. Replay in deep learning: Current approaches and missing biological elements. Neural computation, 33(11):2908–2950, 2021.
  11. The many faces of robustness: A critical analysis of out-of-distribution generalization. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8340–8349, 2021.
  12. Birt: Bio-inspired replay in vision transformers for continual learning. arXiv preprint arXiv:2305.04769, 2023.
  13. Continual learning of a mixed sequence of similar and dissimilar tasks. Advances in Neural Information Processing Systems, 33:18493–18504, 2020.
  14. Introducing language guidance in prompt-based continual learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11463–11473, 2023.
  15. Overcoming catastrophic forgetting in neural networks. Proceedings of the national academy of sciences, 114(13):3521–3526, 2017.
  16. 3d object representations for fine-grained categorization. In Proceedings of the IEEE international conference on computer vision workshops, pages 554–561, 2013.
  17. Learning multiple layers of features from tiny images. 2009.
  18. Overcoming catastrophic forgetting by incremental moment matching. Advances in neural information processing systems, 30, 2017.
  19. The power of scale for parameter-efficient prompt tuning. arXiv preprint arXiv:2104.08691, 2021.
  20. Learn to grow: A continual structure learning framework for overcoming catastrophic forgetting. In International Conference on Machine Learning, pages 3925–3934. PMLR, 2019.
  21. Learning without forgetting. IEEE transactions on pattern analysis and machine intelligence, 40(12):2935–2947, 2017.
  22. Pcr: Proxy-based contrastive replay for online class-incremental continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24246–24255, 2023.
  23. Few-shot class-incremental learning via entropy-regularized data-free replay. In European Conference on Computer Vision, pages 146–162. Springer, 2022.
  24. Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing. ACM Computing Surveys, 55(9):1–35, 2023.
  25. Progressive voronoi diagram subdivision enables accurate data-free class-incremental learning. In The Eleventh International Conference on Learning Representations, 2022.
  26. Class-incremental learning: survey and performance evaluation on image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(5):5513–5533, 2022.
  27. Catastrophic interference in connectionist networks: The sequential learning problem. In Psychology of learning and motivation, pages 109–165. Elsevier, 1989.
  28. An empirical investigation of the role of pre-training in lifelong learning. arXiv preprint arXiv:2112.09153, 2021.
  29. Online class incremental learning on stochastic blurry task boundary via mask and visual prompt tuning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11731–11741, 2023.
  30. Space-time prompting for video class-incremental learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11932–11942, 2023.
  31. Moment matching for multi-source domain adaptation. In Proceedings of the IEEE/CVF international conference on computer vision, pages 1406–1415, 2019.
  32. Effect of scale on catastrophic forgetting in neural networks. In International Conference on Learning Representations, 2021.
  33. Progressive prompts: Continual learning for language models. arXiv preprint arXiv:2301.12314, 2023.
  34. icarl: Incremental classifier and representation learning. In Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, pages 2001–2010, 2017.
  35. Always be dreaming: A new approach for data-free class-incremental learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9374–9384, 2021.
  36. Coda-prompt: Continual decomposed attention-based prompting for rehearsal-free continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11909–11919, 2023.
  37. An introduction to lifelong supervised learning. arXiv preprint arXiv:2207.04354, 2022.
  38. When prompt-based incremental learning does not meet strong pretraining. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 1706–1716, 2023.
  39. Gido M Van de Ven and Andreas S Tolias. Three scenarios for continual learning. arXiv preprint arXiv:1904.07734, 2019.
  40. Beef: Bi-compatible class-incremental learning via energy-based expansion and fusion. In The Eleventh International Conference on Learning Representations, 2022a.
  41. Foster: Feature boosting and compression for class-incremental learning. In European conference on computer vision, pages 398–414, 2022b.
  42. A comprehensive survey of continual learning: Theory, method and application. arXiv preprint arXiv:2302.00487, 2023a.
  43. Hierarchical decomposition of prompt-based continual learning: Rethinking obscured sub-optimality. Advances in Neural Information Processing Systems, 36, 2024.
  44. Isolation and impartial aggregation: A paradigm of incremental learning without interference. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 10209–10217, 2023b.
  45. Dualprompt: Complementary prompting for rehearsal-free continual learning. In European Conference on Computer Vision, pages 631–648. Springer, 2022c.
  46. Learning to prompt for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 139–149, 2022d.
  47. Large scale incremental learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 374–382, 2019.
  48. Der: Dynamically expandable representation for class incremental learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3014–3023, 2021.
  49. Online coreset selection for rehearsal-based continual learning. arXiv preprint arXiv:2106.01085, 2021.
  50. Continual learning through synaptic intelligence. In International conference on machine learning, pages 3987–3995. PMLR, 2017.
  51. Slca: Slow learner with classifier alignment for continual learning on a pre-trained model. arXiv preprint arXiv:2303.05118, 2023.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Zhanxin Gao (1 paper)
  2. Jun Cen (28 papers)
  3. Xiaobin Chang (14 papers)
Citations (13)
View on Semantic Scholar
Github Logo Streamline Icon: https://streamlinehq.com

GitHub

  1. GitHub - Zhanxin-Gao/CPrompt: Consistent Prompting for Rehearsal-Free Continual Learning [CVPR2024] (33 stars)
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets