Research map / Continual learning / Domain and prompt-based
Continual Adaptation of Visual Representations via Domain Randomization and Meta-Learning
Check 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 |
|---|---|---|---|
| Apply heavy image domain randomization during sequential adaptation | Naive fine-tuning · SVHN, P2 | accuracy: 54.0 ± 5.8 -> 72.3 ± 0.9 | Helped |
| Use Meta-DR instead of Naive + DR | Sequential domain adaptation · MNIST-M, P1 | accuracy: 72.0 ± 1.1 -> 75.1 ± 0.5 | Helped |
| Use Meta-DR instead of Naive + DR | Sequential domain adaptation · SVHN, P2 | accuracy: 80.8 ± 0.6 -> 73.8 ± 2.1 | Mixed |
| Add the recall loss (ablation) | Meta-DR · MNIST, P1 | accuracy: 83.7 ± 6.4 -> 94.3 ± 0.7 | Helped |
| Add the recall loss (ablation) | Meta-DR · MNIST-M, P1 | accuracy: 68.8 ± 3.4 -> 76.5 ± 0.6 | Helped |
| Add the adaptation loss (ablation) | Meta-DR · SYN, P1 | accuracy: 92.3 ± 0.4 -> 95.4 ± 0.1 | Helped |
| Add the adaptation loss (ablation) | Meta-DR · SVHN, P1 | accuracy: 86.9 ± 0.1 -> 91.9 ± 0.0 | Helped |
| Combine recall and adaptation losses rather than using adaptation loss alone (ablation) | Meta-DR · MNIST, P1 | accuracy: 89.7 ± 0.5 -> 92.0 ± 0.6 | Helped |
| Combine recall and adaptation losses rather than using recall loss alone (ablation) | Meta-DR · SVHN, P1 | accuracy: 89.5 ± 0.1 -> 91.9 ± 0.2 | Helped |
| Add domain randomization to single-domain training | MNIST-trained model · SVHN domain generalization | accuracy: 23.5 ± 1.6 -> 40.4 ± 1.3 | Helped |
Closest papers by meaning
- Continual Learning of Domain-Invariant Representations (ICML 2026 · similarity 0.88)
- Learning to Continually Learn with the Bayesian Principle (ICML 2024 · similarity 0.86)
- Randomized-MLP Regularization Improves Domain Adaptation and Interpretability in DINOv2 (NeurIPS 2025 · similarity 0.86)
- Random Registers for Cross-Domain Few-Shot Learning (ICML 2025 · similarity 0.86)
- Exploring Sparse Visual Prompt for Domain Adaptive Dense Prediction (AAAI 2024 · similarity 0.85)
- C-LoRA: Continual Low-Rank Adaptation for Pre-trained Visual Models (NeurIPS 2026 · similarity 0.85)
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.