Research map / Learning theory and algorithms / Kernels and regression theory
Learning Taxonomy Adaptation in Large-scale Classification
Rohit Babbar, Ioannis Partalas, Eric Gaussier, Massih-Reza Amini, Cécile Amblard
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 fully hierarchical taxonomy with meta-learned pruned taxonomy (PR-M) | SVM top-down · LSHTC2-3 | error (%): 49.8 -> 49.3 | Improved |
| replace fully hierarchical taxonomy with meta-learned pruned taxonomy (PR-M) | MNB top-down · LSHTC2-4 | error (%): 68.3 -> 65.4 | Improved |
| use top-down hierarchy instead of flat classification | MNB · LSHTC2-3 | error (%): 73.0 -> 62.0 | Improved |
| use top-down hierarchy instead of flat classification | SVM · IPC | error (%): 46.6 -> 51.3 | Hurt |
| use top-down hierarchy instead of flat classification | SVM · LSHTC2-5 | error (%): 54.7 -> 52.7 | Improved |
| prune taxonomy with meta-learned PR-M instead of random pruning | SVM · LSHTC2-4 | error (%): 48.4 -> 47.2 | Improved |
| prune taxonomy with meta-learned PR-M | SVM · IPC | error (%): 46.6 -> 50.7 | Hurt |
| bound-based pruning (PR-B) instead of validation pruning (PR-V) | SVM · LSHTC2-4 | error (%): 46.6 -> 46.5 | No clear effect |
| PR-B instead of earlier perceptron-tree pruning PR-P | SVM · LSHTC2-3 | error (%): 49.6 -> 49.5 | No clear effect |
| use PR-M pruning instead of fully hierarchical | SVM · LSHTC2-5 | Macro-F1: 30.7 -> 31.8 | Improved |
| use PR-M pruning instead of flat | SVM · IPC | Macro-F1: 51.2 -> 48.7 | Hurt |
| use hierarchy instead of flat | SVM · LSHTC2-4 | Macro-F1: 32.9 -> 33.7 | Improved |
Closest papers by meaning
- Deep Taxonomic Networks for Unsupervised Hierarchical Prototype Discovery (NeurIPS 2025 · similarity 0.84)
- Optimal Classification Trees for Continuous Feature Data Using Dynamic Programming with Branch-and-Bound (AAAI 2025 · similarity 0.84)
- Classification Fields: Arbitrarily Fine Recursive Hierarchical Clustering From Few Examples (NeurIPS 2026 · similarity 0.83)
- Happy: A Debiased Learning Framework for Continual Generalized Category Discovery (NeurIPS 2024 · similarity 0.82)
- Expected Probabilistic Hierarchies (NeurIPS 2024 · similarity 0.82)
- Continual Generalized Category Discovery: Learning and Forgetting from a Bayesian Perspective (ICML 2025 · similarity 0.82)
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.
