Research map / Continual learning / Domain and prompt-based
CODA-Prompt: COntinual Decomposed Attention-Based Prompting for Rehearsal-Free Continual Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 CODA-P instead of LwF.MC | continual learning on ImageNet-R · ImageNet-R, 5 tasks | average accuracy (A_N): 73.05 ± 0.50 -> 76.51 ± 0.38 | Helped |
| use CODA-P instead of Deep L2P++ | continual learning on ImageNet-R · ImageNet-R, 10 tasks | average accuracy (A_N): 71.66 ± 0.64 -> 75.45 ± 0.56 | Helped |
| use CODA-P instead of Deep L2P++ | continual learning on ImageNet-R · ImageNet-R, 20 tasks | average accuracy (A_N): 68.42 ± 1.20 -> 72.37 ± 1.19 | Helped |
| use CODA-P instead of Deep L2P++ | continual learning on CIFAR-100 · CIFAR-100, 10 tasks | average accuracy (A_N): 84.30 ± 1.03 -> 86.25 ± 0.74 | Helped |
| use CODA-P instead of LwF.MC | continual learning on DomainNet · DomainNet, 5 tasks | average accuracy (A_N): 74.78 ± 0.43 -> 73.24 ± 0.59 | Hurt |
| use CODA-P instead of DualPrompt | continual learning with dual semantic and covariate shifts · ImageNet-R with covariate domain shifts, 5 tasks | average accuracy (A_N): 66.98 ± 0.08 -> 71.35 ± 0.08 | Helped |
| add attention to prompt-component weighting (ablation) | CODA-P · ImageNet-R, 10 tasks | average accuracy (A_N): 74.52 ± 0.65 -> 75.45 ± 0.56 | Helped |
| freeze past task components (ablation) | CODA-P · ImageNet-R, 10 tasks | average accuracy (A_N): 74.60 ± 0.64 -> 75.45 ± 0.56 | Helped |
| add orthogonality regularization (ablation) | CODA-P · ImageNet-R, 10 tasks | average accuracy (A_N): 70.66 ± 0.60 -> 75.45 ± 0.56 | Helped |
Closest papers by meaning
- REP: Resource-Efficient Prompting for Rehearsal-Free Continual Learning (NeurIPS 2025 · similarity 0.92)
- Convolutional Prompting meets Language Models for Continual Learning (CVPR 2024 · similarity 0.91)
- Evolving Parameterized Prompt Memory for Continual Learning (AAAI 2024 · similarity 0.90)
- Mixture of Experts Meets Prompt-Based Continual Learning (NeurIPS 2024 · similarity 0.90)
- Visual Prompt Tuning in Null Space for Continual Learning (NeurIPS 2024 · similarity 0.89)
- Consistent Prompting for Rehearsal-Free Continual Learning (CVPR 2024 · similarity 0.88)
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.