Research map / Continual learning / Class-incremental
PANDA – Patch and Distribution-Aware Augmentation for Long-Tailed Exemplar-Free Continual Learning
Siddeshwar Raghavan, Jiangpeng He, Fengqing Zhu
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 PANDA augmentation to PTM-based EFCL methods | CodaPrompt · CIFAR100-LT (rho=0.01) | average accuracy: 76.52 -> 87.49 | Improved |
| add PANDA augmentation | RanPAC · CIFAR100-LT (rho=0.01) | average accuracy: 90.35 -> 91.91 | Improved |
| add PANDA augmentation | MOS · CIFAR100-LT (rho=0.01) | average accuracy: 91.60 -> 92.04 | Improved |
| add PANDA augmentation | CoFiMA · iNaturalist (100 classes) | average accuracy: 94.55 -> 93.56 | Hurt |
| add PANDA augmentation | CoFiMA · iNaturalist (100 classes) | average forgetting: 3.88 -> 2.98 | Improved |
| add PANDA augmentation | FeCAM · iNaturalist (100 classes) | average forgetting: 3.33 -> 4.44 | Hurt |
| add PANDA augmentation under dual-level imbalance | CodaPrompt · CIFAR100-LT dual-level (task 3) | average accuracy: 82.88 -> 87.77 | Improved |
| add PANDA augmentation under dual-level imbalance | MOS · CIFAR100-LT dual-level (task 2) | average accuracy: 93.54 -> 92.22 | Hurt |
| add PANDA augmentation under dual-level imbalance | MOS · CIFAR100-LT dual-level (tasks 3 and 4) | average accuracy: 92.10 -> 93.21; 91.69 -> 92.82 | Improved |
| use PANDA instead of CutMix/Mixup/Remix/Con-CutMix augmentation (ablation) | RanPAC · CIFAR100-LT (rho=0.02) | average accuracy: 87.27 -> 90.31 (vs Con-CutMix) | Improved |
| use PANDA instead of Contrastive CutMix under dual-level imbalance (ablation) | RanPAC · CIFAR100-LT dual-level (rho=0.02) | average accuracy: 84.19 -> 90.08 | Improved |
| CLIP-similarity patch selection instead of attention-affinity masks (ablation) | APART · CIFAR100-LT (rho=0.01) | average accuracy: 80.87 -> 83.39 | Improved |
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.
- RanPAC — RanPAC: Random Projections and Pre-trained Models for Continual Learning (NeurIPS 2023) · better on 4 of 4 shared results
- FeCAM — FeCAM: Exploiting the Heterogeneity of Class Distributions in Exemplar-Free Continual Learning (NeurIPS 2023) · 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
- Online Task-Free Continual Learning via Dynamic Expansionable Memory Distribution (CVPR 2025 · similarity 0.88)
- LoRA Subtraction for Drift-Resistant Space in Exemplar-Free Continual Learning (CVPR 2025 · similarity 0.87)
- Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning (NeurIPS 2025 · similarity 0.87)
- SAFE: Slow and Fast Parameter-Efficient Tuning for Continual Learning with Pre-Trained Models (NeurIPS 2024 · similarity 0.87)
- Long-Tail Class Incremental Learning via Independent Sub-prototype Construction (CVPR 2024 · similarity 0.87)
- PaLoRA: Paced Low-Rank Adaptation for Continual Learning (NeurIPS 2026 · 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.
