Research map / Label-efficient and robust learning / Recognition and long-tail
HyperMiner: Topic Taxonomy Mining with Hyperbolic Embedding
Yi.shi Xu, Dongsheng Wang, Bo Chen, Ruiying Lu, Zhibin Duan, Mingyuan Zhou
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 Euclidean inner-product similarity with hyperbolic distance in SawETM (HyperMiner) (ablation) | SawETM · 20NG | km-Purity: 43.36 -> 44.37 | Improved |
| replace Euclidean with hyperbolic distance in ETM (HyperETM) (ablation) | ETM · 20NG | km-NMI: 37.72 -> 39.06 | Improved |
| hyperbolic embeddings in ETM at low embedding dimension (ablation) | ETM · 20NG classification D=2 | accuracy: 19.87 -> 24.33 | Improved |
| hyperbolic embeddings in ETM at D=50 (ablation) | ETM · 20NG classification D=50 | accuracy: 43.85 -> 44.38 | Improved |
| hyperbolic embeddings in SawETM at D=2 (ablation) | SawETM · 20NG classification D=2 | accuracy: 16.74 -> 20.16 | Improved |
| add hyperbolic contrastive prior-knowledge loss (HyperMiner-KG) (ablation) | HyperMiner · 20NG | km-Purity: 44.37 -> 45.16 | Improved |
| add hyperbolic contrastive prior-knowledge loss (HyperMiner-KG) (ablation) | HyperMiner · TMN | km-NMI: 41.93 -> 42.81 | Improved |
| increase regularization coefficient lambda from 0.1 to 5 (ablation) | HyperMiner-KG · 20NG | km-Purity: 43.76 -> 45.16 | Improved |
| increase regularization coefficient lambda from 5 to 10 (ablation) | HyperMiner-KG · 20NG | km-Purity: 45.16 -> 44.81 | Hurt |
| use TopicNet Gaussian knowledge guidance instead of none (vs SawETM base) | SawETM · 20NG | km-Purity: 43.36 -> 42.94 | Hurt |
Later papers that compared with this method
1 later accepted paper reports HyperMiner as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Effective Neural Topic Modeling with Embedding Clustering Regularization (ICML 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
- Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry (ICML 2018 · similarity 0.87)
- Knowledge-Aware Bayesian Deep Topic Model (NeurIPS 2022 · similarity 0.87)
- Representation Tradeoffs for Hyperbolic Embeddings (ICML 2018 · similarity 0.86)
- Hyperbolic Embeddings of Supervised Models (NeurIPS 2024 · similarity 0.84)
- Low-distortion and GPU-compatible Tree Embeddings in Hyperbolic Space (ICML 2025 · similarity 0.83)
- Learning Structured Representations with Hyperbolic Embeddings (NeurIPS 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.
