Research map / Continual learning / Class-incremental
DiffClass: Diffusion-Based Class Incremental Learning
Zichong Meng, Jie Zhang, Changdi Yang, Zheng Zhan, Pu Zhao, Yanzhi Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace baseline synthetic-data CIL with DiffClass | exemplar-free CIL on CIFAR100 · CIFAR100 N=20 | Acc_L: 42.87 -> 57.11 vs SEED | Improved |
| replace SEED with DiffClass | exemplar-free CIL on ImageNet100 · ImageNet100 N=20 | Acc_L: 45.77 -> 68.68 | Improved |
| add multi-distribution matching (MDM) diffusion fine-tuning (ablation) | DiffClass · ImageNet100 N=5 | Acc_avg: 59.71 -> 74.85 | Improved |
| add multi-domain adaptation (domain classifier with gradient reversal) (ablation) | DiffClass · ImageNet100 N=5 | Acc_avg: 65.29 -> 74.85 | Improved |
| add selective synthetic image augmentation (SSIA) (ablation) | DiffClass · ImageNet100 N=5 | Acc_avg: 62.37 -> 74.85 | Improved |
| train with synthetic previous-task data instead of real data | ResNet with linear classifier · ImageNet100 20-task diagnosis, first four tasks | total accuracy: 82.72 -> 59.37 | Hurt |
| increase number of tasks N from 10 to 20 | SEED · CIFAR100 | Acc_L: 51.42 -> 42.87 | Hurt |
| increase number of tasks N from 10 to 20 | DiffClass · CIFAR100 | Acc_L: 58.40 -> 57.11 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- R-DFCIL — R-DFCIL: Relation-Guided Representation Learning for Data-Free Class Incremental Learning (ECCV 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental Learning (ICLR 2023 · similarity 0.90)
- Make Domain Shift a Catastrophic Forgetting Alleviator in Class-Incremental Learning (AAAI 2025 · similarity 0.90)
- Generative Multi-modal Models are Good Class Incremental Learners (CVPR 2024 · similarity 0.89)
- Defying Imbalanced Forgetting in Class Incremental Learning (AAAI 2024 · similarity 0.88)
- Gradual Divergence for Seamless Adaptation: A Novel Domain Incremental Learning Method (ICML 2024 · similarity 0.88)
- Non-exemplar Online Class-Incremental Continual Learning via Dual-Prototype Self-Augment and Refinement (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
