Research map / Continual learning / Class-incremental
Class-Incremental Learning with Cross-Space Clustering and Controlled Transfer
Arjun Ashok, K J Joseph, Vineeth N Balasubramanian
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 CSCCT to iCaRL | iCaRL · CIFAR100 (B=C=1) | average incremental accuracy (%): 30.92 -> 34.02 | Improved |
| add CSCCT to LUCIR | LUCIR · CIFAR100 (B=50, C=2) | average incremental accuracy (%): 55.38 -> 56.49 | Improved |
| add CSCCT to PODNet | PODNet · CIFAR100 (B=C=1) | average incremental accuracy (%): 33.58 -> 36.23 | Improved |
| add CSCCT to PODNet | PODNet · ImageNet-Subset (B=50, C=10) | average incremental accuracy (%): 75.32 -> 76.41 | Improved |
| add CSCCT to LUCIR | LUCIR · CIFAR100 (B=C=1) | average incremental accuracy (%): 25.4 -> 28.12 | Improved |
| add Cross-Space Clustering alone (ablation) | LUCIR · CIFAR100 (B=50, C=1) | average incremental accuracy (%): 50.26 -> 52.04 | Improved |
| add Controlled Transfer alone (ablation) | LUCIR · CIFAR100 (B=50, C=1) | average incremental accuracy (%): 50.26 -> 51.5 | Improved |
| add Controlled Transfer alone (ablation) | LUCIR · CIFAR100 (B=50, C=5) | average incremental accuracy (%): 59.4 -> 61.97 | Improved |
| add Cross-Space Clustering alone (ablation) | LUCIR · CIFAR100 (B=50, C=5) | average incremental accuracy (%): 59.4 -> 60.45 | Improved |
| combine CSC and CT (full CSCCT) instead of CSC alone (ablation) | LUCIR + CSC · CIFAR100 (B=C=1) | average incremental accuracy (%): 27.16 -> 28.12 | Improved |
| increase batch size from 128 to 1024 (ablation) | LUCIR · CIFAR100 (B=50, C=5) | average incremental accuracy (%): 59.4 -> 54.83 | Hurt |
| increase batch size from 128 to 1024 (ablation) | LUCIR + CSCCT · CIFAR100 (B=50, C=5) | average incremental accuracy (%): 60.01 -> 57.8 | 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 4 of 4 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
- Few-Shot Class-Incremental Learning via Class-Aware Bilateral Distillation (CVPR 2023 · similarity 0.90)
- Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · similarity 0.90)
- Reducing Class-wise Confusion for Incremental Learning with Disentangled Manifolds (CVPR 2025 · similarity 0.89)
- Long-Tail Class Incremental Learning via Independent Sub-prototype Construction (CVPR 2024 · similarity 0.89)
- Adaptive Discovering and Merging for Incremental Novel Class Discovery (AAAI 2024 · similarity 0.88)
- Order-Robust Class Incremental Learning: Graph-Driven Dynamic Similarity Grouping (CVPR 2025 · similarity 0.88)
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.
