Research map / Continual learning / Class-incremental
Learnability and Algorithm for Continual Learning
Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, Bing Liu
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 |
|---|---|---|---|
| add a separate WP head on top of the OOD-head model (ROW vs ROW (-WP)) (ablation) | ROW · C10-5T, C100-10T, C100-20T | average classification accuracy: +2.47, +2.43, +2.63 | Improved |
| add distance-based (Mahalanobis) coefficient to TP probability (ROW (-WP) vs ROW (-WP-MD)) (ablation) | ROW (-WP) · C10-5T, C100-10T, C100-20T | average classification accuracy: +4.44, +4.76, +6.12 | Improved |
| halve replay memory buffer size (ablation) | ROW · Average over 5 benchmarks | average classification accuracy: -1.02 avg (73.72 to 72.70) | No clear effect |
| halve replay memory buffer size (ablation) | DER++ · Average over 5 benchmarks | average classification accuracy: -4.73 avg (66.89 to 62.16) | Hurt |
| ROW (full method with OOD replay and WP head) vs replay baseline MORE | ROW · Average over 5 benchmarks | average classification accuracy: +2.13 (73.72 vs 71.59) | Improved |
| replace pre-trained feature extractor with adapter-tuned DeiT (vs frozen SLDA classifier-only) | SLDA · C100-10T, T-5T | average classification accuracy: SLDA 67.82 / 57.93 vs ROW 74.72 / 65.11 | Hurt |
| remove classes overlapping with CL datasets from pre-training (L2P) | L2P · Average over 5 benchmarks | average classification accuracy: drops to 60.47 average | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- iCaRL — iCaRL: Incremental Classifier and Representation Learning (CVPR 2017) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Class Incremental Learning via Likelihood Ratio Based Task Prediction (ICLR 2024 · similarity 0.91)
- Continual Learning Using a Kernel-Based Method Over Foundation Models (AAAI 2025 · similarity 0.89)
- eTag: Class-Incremental Learning via Embedding Distillation and Task-Oriented Generation (AAAI 2024 · similarity 0.88)
- AnaCP: Toward Upper-Bound Continual Learning via Analytic Contrastive Projection (NeurIPS 2025 · similarity 0.88)
- GACL: Exemplar-Free Generalized Analytic Continual Learning (NeurIPS 2024 · similarity 0.88)
- Provable Contrastive Continual Learning (ICML 2024 · 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.
