Research map / Continual learning / Class-incremental
Two-Way Is Better Than One: Bidirectional Alignment with Cycle Consistency for Exemplar-Free Class-Incremental Learning
Hongye Xu, Bartosz Krawczyk
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 AdaGauss two-stage drift compensation with BiCyc bidirectional cycle-consistent alignment | AdaGauss · CIFAR-100 T=10 (from scratch) | A_last: 46.8 -> 50.6 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · CIFAR-100 T=20 (from scratch) | A_inc: 54.4 -> 56.5 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · TinyImageNet T=20 (from scratch) | A_inc: 39.5 -> 44.2 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · TinyImageNet T=10 (from scratch) | A_inc: 45.8 -> 49.1 | Improved |
| replace DPCR with BiCyc | DPCR · CIFAR-100 T=10 (from scratch) | A_inc: 62.8 -> 63.2 | Improved |
| replace DPCR with BiCyc | DPCR · ImageNet-100 T=20 (from scratch) | A_inc: 54.7 -> 58.2 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · ImageNet-100 T=10 (from scratch) | A_inc: 65.0 -> 66.8 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · CUB-200 T=20 (ImageNet-pretrained) | A_inc: 57.0 -> 55.9 | Hurt |
| replace AdaGauss with BiCyc | AdaGauss · CUB-200 T=20 (ImageNet-pretrained) | A_last: 45.0 -> 43.7 | Hurt |
| replace AdaGauss with BiCyc | AdaGauss · CUB-200 T=10 (ImageNet-pretrained) | A_inc: 63.4 -> 64.0 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · CIFAR-100 T=10 (from scratch) | F_last forgetting: 16.7 -> 13.5 | Improved |
| replace AdaGauss with BiCyc | AdaGauss · TinyImageNet T=20 (from scratch) | F_last forgetting: 23.1 -> 18.9 | Improved |
Closest papers by meaning
- Navigating Semantic Drift in Task-Agnostic Class-Incremental Learning (ICML 2025 · similarity 0.88)
- Non-exemplar Online Class-Incremental Continual Learning via Dual-Prototype Self-Augment and Refinement (AAAI 2024 · similarity 0.88)
- Semantic Shift Estimation via Dual-Projection and Classifier Reconstruction for Exemplar-Free Class-Incremental Learning (ICML 2025 · similarity 0.87)
- Task-recency bias strikes back: Adapting covariances in Exemplar-Free Class Incremental Learning (NeurIPS 2024 · similarity 0.87)
- Adapter Merging with Centroid Prototype Mapping for Scalable Class-Incremental Learning (CVPR 2025 · similarity 0.87)
- Elastic Feature Consolidation For Cold Start Exemplar-Free Incremental Learning (ICLR 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.
