Research map / Continual learning / Class-incremental
Continual Learning with Evolving Class Ontologies
Zhiqiu Lin, Deepak Pathak, Yu-Xiong Wang, Deva Ramanan, Shu Kong
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 |
|---|---|---|---|
| finetune previous-TP model instead of training from scratch | LabelNew · iNat-LECO | mAcc: 65.69 -> 73.64 | Improved |
| freeze the previous-TP backbone instead of training from scratch | LabelNew · iNat-LECO | mAcc: 65.69 -> 50.82 | Hurt |
| relabel old data instead of labeling new data | FinetunePrev · iNat-LECO | mAcc: 73.64 -> 71.26 | No clear effect |
| add ST-Soft semi-supervised loss on old data | FinetunePrev LabelNew · iNat-LECO | mAcc: 73.64 -> 79.64 | Improved |
| add joint training loss using old coarse labels | FinetunePrev LabelNew · iNat-LECO | mAcc: 73.64 -> 82.98 | Improved |
| combine Joint and ST-Soft losses | FinetunePrev LabelNew · iNat-LECO | mAcc: 73.64 -> 83.61 | Improved |
| add Joint loss on top of ST-Soft | ST-Soft · CIFAR-LECO | mAcc: 69.86 -> 69.77 | No clear effect |
| add LPL loss using taxonomic hierarchy | FinetunePrev LabelNew · iNat-LECO | mAcc: 73.64 -> 83.62 | Improved |
| add SSL with conditioning pseudo-label refinement to LPL (ST-Hard) | LPL · iNat-LECO | mAcc: 83.62 -> 84.34 | Improved |
| add SSL with filtering refinement to LPL (ST-Hard) | LPL · iNat-LECO | mAcc: 83.62 -> 84.23 | Improved |
| add SSL/Cond to LPL (ST-Hard) | LPL · CIFAR-LECO | mAcc: 69.31 -> 70.42 | Improved |
| use Conditioning instead of Filtering with FixMatch (ablation) | LPL + FixMatch · iNat-4TP-LECO TP3 | mAcc: 85.3 -> 78.8 | Hurt |
Closest papers by meaning
- CLEO: Continual Learning of Evolving Ontologies (ECCV 2024 · similarity 0.89)
- Preparing the Future for Continual Semantic Segmentation (ICCV 2023 · similarity 0.85)
- Learning Visually Grounded Domain Ontologies via Embodied Conversation and Explanation (AAAI 2025 · similarity 0.84)
- Hierarchical Semantic Tree Anchoring for CLIP-Based Class-Incremental Learning (NeurIPS 2026 · similarity 0.84)
- Enhancing Visual Continual Learning with Language-Guided Supervision (CVPR 2024 · similarity 0.83)
- Adaptive Discovering and Merging for Incremental Novel Class Discovery (AAAI 2024 · similarity 0.83)
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.
