Research map / Continual learning / Online and replay-based
Generalized Variational 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 |
|---|---|---|---|
| Add task-specific FiLM layers to GVCL (ablation) | GVCL · Easy-CHASY | ACC (%): $2.0 \pm 0.5$\% | Helped |
| Add task-specific FiLM layers to GVCL (ablation) | GVCL · Hard-CHASY | ACC (%): $5.1 \pm 0.4$\% | Helped |
| Add task-specific FiLM layers to GVCL (ablation) | GVCL · Split-MNIST (10 tasks) | ACC (%): $4.0 \pm 0.7$\% | Helped |
| Add task-specific FiLM layers to GVCL (ablation) | GVCL · Split-CIFAR | ACC (%): $9.5 \pm 1.4$\% | Helped |
| Add task-specific FiLM layers to GVCL (ablation) | GVCL · Mixed Vision Tasks | ACC (%): $31.0 \pm 2.2$\% | Helped |
| Add task-specific FiLM layers to GVCL (ablation) | GVCL · Average across benchmarks | ACC (%): $10.3 \pm 10.6$\% | Helped |
| Add task-specific FiLM layers to variational methods | GVCL-F versus GVCL · Mixed Vision Tasks | Expected Calibration Error (%): 0.32\% compared to HAT's 1.69\% | Helped |
| Use GVCL-F instead of the strongest reported baseline | Continual learning models · Easy-CHASY | ACC (%): 82.6 \pm 0.9 -> 90.9 \pm 0.3 | Helped |
| Use GVCL-F instead of the strongest reported baseline | Continual learning models · Hard-CHASY | ACC (%): 64.8 \pm 0.8 -> 69.5 \pm 0.6 | Helped |
| Use GVCL-F instead of the strongest reported baseline | Continual learning models · Split-MNIST (10 Tasks) | ACC (%): 98.3 \pm 0.1 -> 98.6 \pm 0.1 | Helped |
| Use GVCL-F instead of the strongest reported baseline | Continual learning models · Split-CIFAR | ACC (%): 77.3 \pm 0.3 -> 80.0 \pm 0.5 | Helped |
| Use GVCL-F instead of the strongest reported baseline | Continual learning models · Mixed Vision Tasks | ACC (%): 80.3 \pm 1.0 -> 80.0 \pm 1.2 | No clear effect |
Closest papers by meaning
- Variational Continual Learning (ICLR 2018 · similarity 0.93)
- A Statistical Theory of Regularization-Based Continual Learning (ICML 2024 · similarity 0.89)
- Temporal-Difference Variational Continual Learning (NeurIPS 2025 · similarity 0.88)
- GACL: Exemplar-Free Generalized Analytic Continual Learning (NeurIPS 2024 · similarity 0.88)
- Optimal Rates in Continual Linear Regression via Increasing Regularization (NeurIPS 2025 · similarity 0.87)
- Learning Continually by Spectral Regularization (ICLR 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.