Research map / Continual learning / Class-incremental
EFC++: Elastic Feature Consolidation with Prototype Re-balancing for Cold Start Exemplar-free Incremental Learning
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 PR-ACE joint training with decoupled training plus post-training prototype re-balancing (EFC to EFC++) (ablation) | EFC · CIFAR-100 (Cold Start, 10 step) | final per-step accuracy: 43.62 -> 47.52 | Improved |
| add prototype re-balancing in place of PR-ACE (EFC to EFC++) (ablation) | EFC · ImageNet-Subset (Cold Start, 10 step) | final per-step accuracy: 47.38 -> 53.90 | Improved |
| add prototype re-balancing in place of PR-ACE (EFC to EFC++) (ablation) | EFC · Tiny-ImageNet (Cold Start, 20 step) | final per-step accuracy: 28.69 -> 32.56 | Improved |
| add prototype re-balancing in place of PR-ACE (EFC to EFC++) (ablation) | EFC · CIFAR-100 (Warm Start, 10 step) | final per-step accuracy: 60.87 -> 62.15 | Improved |
| add prototype re-balancing in place of PR-ACE (EFC to EFC++) (ablation) | EFC · ImageNet-1K (Warm Start, 10 step) | final per-step accuracy: 59.49 -> 59.09 | No clear effect |
| add prototype re-balancing in place of PR-ACE (EFC to EFC++) (ablation) | EFC · ImageNet-1K (Cold Start, 10 step) | final per-step accuracy: 42.62 -> 44.72 | Improved |
| add EFM-based update of prototype means (ablation) | EFC++ · Tiny-ImageNet (Cold Start, 20 step) | last-step accuracy: 29.04 -> 32.56 | Improved |
| add EFM-based update of prototype means (ablation) | EFC++ · ImageNet-Subset (Cold Start, 10 step) | last-step accuracy: 52.95 -> 53.90 | Improved |
| use real class means instead of EFM-estimated means (ablation) | EFC++ · ImageNet-Subset (Cold Start, 20 step) | last-step accuracy: 40.86 -> 49.70 | Improved |
| use real covariance instead of fixed covariance (with real means) (ablation) | EFC++ · ImageNet-Subset (Cold Start, 10 step) | last-step accuracy: 56.14 -> 55.69 | No clear effect |
| use real covariance instead of fixed covariance (with EFM means) (ablation) | EFC++ · ImageNet-Subset (Cold Start, 10 step) | last-step accuracy: 53.90 -> 52.44 | Hurt |
| use real covariance instead of fixed covariance (with real means) (ablation) | EFC++ · Tiny-ImageNet (Cold Start, 10 step) | last-step accuracy: 39.94 -> 39.95 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- FeCAM — FeCAM: Exploiting the Heterogeneity of Class Distributions in Exemplar-Free Continual Learning (NeurIPS 2023) · better on 4 of 4 shared results
- R-DFCIL — R-DFCIL: Relation-Guided Representation Learning for Data-Free Class Incremental Learning (ECCV 2022) · better on 2 of 2 shared results
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
- Elastic Feature Consolidation For Cold Start Exemplar-Free Incremental Learning (ICLR 2024 · similarity 0.98)
- Task-recency bias strikes back: Adapting covariances in Exemplar-Free Class Incremental Learning (NeurIPS 2024 · similarity 0.90)
- Data-Free Reservoir Features for Efficient Long-Horizon Cold-Start Continual Learning (NeurIPS 2026 · similarity 0.88)
- Semantic Shift Estimation via Dual-Projection and Classifier Reconstruction for Exemplar-Free Class-Incremental Learning (ICML 2025 · similarity 0.87)
- Resurrecting Old Classes with New Data for Exemplar-Free Continual Learning (CVPR 2024 · similarity 0.86)
- eTag: Class-Incremental Learning via Embedding Distillation and Task-Oriented Generation (AAAI 2024 · similarity 0.86)
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.
