Research map / Continual learning / Online and replay-based
AttriCLIP: A Non-Incremental Learner for Incremental Knowledge Learning
Runqi Wang, Xiaoyue Duan, Guoliang Kang, Jianzhuang Liu, Shaohui Lin, Songcen Xu, Jinhu Lü, Baochang Zhang
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 |
|---|---|---|---|
| use attribute-keyed prompts rather than category-specific prompt learning | AttriCLIP · CIFAR100, Task10 | average accuracy: 79.7 vs 66.7 | Improved |
| use attribute-keyed prompts rather than category-specific prompt learning | AttriCLIP · ImageNet100, Task10 | average accuracy: 81.4 vs 75.4 | Improved |
| transfer learned attribute prompts from ImageNet100 before training on CIFAR100 | AttriCLIP · CIFAR100-I2C | forward transfer: +0.9 | Improved |
| transfer learned attribute prompts from CIFAR100 to ImageNet100 | AttriCLIP · ImageNet100-I2C | backward transfer: +7.0 | Improved |
| train with attribute-keyed prompts for the sequential ImageNet100 and CIFAR100 setting | AttriCLIP · CIFAR100 + ImageNet100 CDCL | accuracy: 78.3 vs 67.1 | Improved |
| use cosine distance rather than triplet loss for key matching (ablation) | AttriCLIP · CIFAR100 | average accuracy: 80.22 -> 81.38 | Improved |
| use cosine distance rather than MSE for key matching (ablation) | AttriCLIP · CIFAR100 | average accuracy: 80.81 -> 81.38 | Improved |
| add prompt diversity loss with λp=0.3 instead of omitting it (ablation) | AttriCLIP · CIFAR100 | average accuracy: 78.10 -> 81.38 | Improved |
| set key matching loss weight λk to 0.7 (ablation) | AttriCLIP · CIFAR100 | average accuracy: 80.28 -> 81.38 | Improved |
| vary prompt length, attribute-bank size, and selected-key count (ablation) | AttriCLIP · CIFAR100 | average accuracy | Mixed |
Later papers that compared with this method
2 later accepted papers report AttriCLIP as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Gradient Projection for Continual Parameter-Efficient Tuning (TPAMI 2025 · better on 1 of 1 shared result)
- Semantic Residual Prompts for Continual Learning (ECCV 2024 · better on 2 of 2 shared results)
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.
- DualPrompt-1 — DualPrompt: Complementary Prompting for Rehearsal-Free Continual Learning (ECCV 2022) · better on 1 of 1 shared result
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
- Boosting Domain Incremental Learning: Selecting the Optimal Parameters is All You Need (CVPR 2025 · similarity 0.82)
- K-DeCore: Facilitating Knowledge Transfer in Continual Structured Knowledge Reasoning via Knowledge Decoupling (NeurIPS 2025 · similarity 0.81)
- Knowledge Graph Enhanced Generative Multi-modal Models for Class-Incremental Learning (NeurIPS 2025 · similarity 0.80)
- Fine-Grained Knowledge Selection and Restoration for Non-exemplar Class Incremental Learning (AAAI 2024 · similarity 0.80)
- Long-Tail Class Incremental Learning via Independent Sub-prototype Construction (CVPR 2024 · similarity 0.80)
- Knowledge Memorization and Rumination for Pre-trained Model-based Class-Incremental Learning (CVPR 2025 · similarity 0.79)
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.
