Research map / Label-efficient and robust learning / Recognition and long-tail
Dual-Margin Embedding for Fine-Grained Long-Tailed Plant Taxonomy
Cheng-Yaw Low, Heejoon Koo, Jaewoo Park, Meeyoung Cha
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 dual-margin penalization over base margin only (ablation) | TaxoNet · AA-Central | macro recall: 63.91 -> 67.59 | Improved |
| add norm-guided oversampling over no oversampling (ablation) | TaxoNet (dual-margin) · AA-Central | macro recall: 67.59 -> 69.48 | Improved |
| norm-guided instead of random oversampling selection (ablation) | TaxoNet (dual-margin) · AA-Central | macro recall: 67.56 -> 69.48 | Improved |
| random-selection oversampling instead of none (ablation) | TaxoNet (dual-margin) · AA-Central | macro recall: 67.59 -> 67.56 | No clear effect |
| add power-based regularization (ablation) | TaxoNet (B, E) · AA-Central | macro recall: 69.48 -> 72.90 | Improved |
| add power-based regularization (ablation) | TaxoNet (B, E) · iNat-Plantae | macro recall: 81.87 -> 83.21 | Improved |
| add power-based regularization (ablation) | TaxoNet (B, E) · AA-East | rank-1 accuracy: 84.53 -> 84.32 | Hurt |
| TaxoNet versus LDAM, strongest baseline | AA-Central · AA-Central | rank-1 accuracy: 92.31 -> 91.92 | Hurt |
| TaxoNet versus LDAM on tail genus classes | AA-Central · AA-Central genus tail | macro recall: 44.18 -> 57.06 | Improved |
| TaxoNet versus LDAM on head genus classes | AA-Central · AA-Central genus head | macro recall: 94.13 -> 92.37 | Hurt |
| TaxoNet versus LDAM under regional shift | AA-Central model on AA-East · AA-East transfer | rank-1 accuracy: 49.26 -> 52.13 | Improved |
| TaxoNet versus LDAM in open-set rejection | AA-Central · Open-set AA-Central | TNR: 88.59 -> 91.28 | Improved |
Closest papers by meaning
- Open-Insect: Benchmarking Open-Set Recognition of Novel Species in Biodiversity Monitoring (NeurIPS 2025 · similarity 0.86)
- Deep Taxonomic Networks for Unsupervised Hierarchical Prototype Discovery (NeurIPS 2025 · similarity 0.86)
- BioTrove: A Large Curated Image Dataset Enabling AI for Biodiversity (NeurIPS 2024 · similarity 0.85)
- DivShift: Exploring Domain-Specific Distribution Shifts in Large-Scale, Volunteer-Collected Biodiversity Datasets (AAAI 2025 · similarity 0.84)
- BioCLIP: A Vision Foundation Model for the Tree of Life (CVPR 2024 · similarity 0.84)
- OAM-TCD: A globally diverse dataset of high-resolution tree cover maps (NeurIPS 2024 · similarity 0.84)
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.
