Research map / Continual learning / Online and replay-based
Uncertainty-guided Continual Learning with Bayesian Neural Networks
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 uncertainty to adapt per-parameter learning rates | Bayesian neural networks · 5-Split MNIST | ACC (%): 99.63 vs 99.59 | Helped |
| Use uncertainty to adapt per-parameter learning rates | Bayesian neural networks · Permuted MNIST, 1.9M parameters | ACC (%): 97.42 vs 97.34 | Helped |
| Use uncertainty-guided learning-rate regularization instead of HAT | Bayesian neural networks · Alternating CIFAR10/100 | ACC (%): 79.44 vs 78.32 | Helped |
| Use uncertainty-guided learning-rate regularization instead of HAT | Bayesian neural networks · Alternating CIFAR10/100 | BWT (%): -0.72 vs -0.04 | Hurt |
| Use uncertainty-guided learning-rate regularization instead of HAT | Bayesian neural networks · Sequence of 8 tasks | ACC (%): 84.04 vs 81.59 | Helped |
| Use uncertainty-guided learning-rate regularization instead of HAT | Bayesian neural networks · Sequence of 8 tasks | BWT (%): -0.84 vs -0.14 | Hurt |
| Regularize posterior means only, using importance 1/σ rather than scales or both (ablation) | UCB · 2-Split MNIST | ACC (%): 98.0 -> 99.2 | Helped |
| Use importance 1/σ instead of |μ|/σ for mean regularization (ablation) | UCB · 2-Split MNIST | ACC (%): 98.4 -> 99.2 | Helped |
| Use 10 rather than 1 Monte Carlo sample (ablation) | UCB · 2-Split MNIST | ACC (%): 98.0 -> 99.2 | Helped |
| Increase Monte Carlo samples from 10 to 15 (ablation) | UCB · 2-Split MNIST | ACC (%): 99.2 -> 98.3 | Hurt |
Closest papers by meaning
- Model Sensitivity Aware Continual Learning (NeurIPS 2024 · similarity 0.89)
- Learning to Continually Learn with the Bayesian Principle (ICML 2024 · similarity 0.89)
- Model Uncertainty Quantification by Conformal Prediction in Continual Learning (ICML 2025 · similarity 0.88)
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.88)
- TinySubNets: An Efficient and Low Capacity Continual Learning Strategy (AAAI 2025 · similarity 0.88)
- BECAME: Bayesian Continual Learning with Adaptive Model Merging (ICML 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.