Research map / Continual learning / Domain and prompt-based
A Unified Approach to Domain Incremental Learning with Memory: Theory and Algorithm
Haizhou Shi, Hao 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 fixed replay coefficients (DER++) with adaptive learned coefficients and domain alignment (UDIL) | DER++ · R-MNIST | average accuracy: 84.258 -> 86.635 | Improved |
| adaptive coefficients vs DER++ on R-MNIST forgetting | DER++ · R-MNIST | forgetting: 13.692 -> 8.506 | Improved |
| adaptive coefficients vs best baseline on Seq-CORe50 | DER++ · Seq-CORe50 | average accuracy: 78.629 -> 82.103 | Improved |
| adaptive coefficients vs DER++ on Seq-CORe50 forward transfer | DER++ · Seq-CORe50 | forward transfer: 22.488 -> 31.215 | Improved |
| adaptive coefficients vs CLS-ER on HD-Balls | CLS-ER · HD-Balls | average accuracy: 85.844 -> 86.872 | Improved |
| UDIL vs DER++ on permutation MNIST accuracy | DER++ · P-MNIST | average accuracy: 92.950 -> 92.666 | Hurt |
| UDIL vs DER++ on permutation MNIST forgetting | DER++ · P-MNIST | forgetting: 3.378 -> 2.853 | Improved |
| add memory replay buffer (ER) to fine-tuning | Fine-tune · Seq-CORe50 | average accuracy: 31.832 -> 66.605 | Improved |
Closest papers by meaning
- Complementary Domain Adaptation and Generalization for Unsupervised Continual Domain Shift Learning (ICCV 2023 · similarity 0.89)
- On Generalizing Beyond Domains in Cross-Domain Continual Learning (CVPR 2022 · similarity 0.88)
- Prototype-Guided Continual Adaptation for Class-Incremental Unsupervised Domain Adaptation (ECCV 2022 · similarity 0.88)
- Advancing Cross-domain Discriminability in Continual Learning of Vision-Language Models (NeurIPS 2024 · similarity 0.88)
- Make Domain Shift a Catastrophic Forgetting Alleviator in Class-Incremental Learning (AAAI 2025 · similarity 0.87)
- Boosting Domain Incremental Learning: Selecting the Optimal Parameters is All You Need (CVPR 2025 · 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.
