Research map / Label-efficient and robust learning / Recognition and long-tail
Ontologue: Declarative Benchmark Construction for Ontological Multi-Label Classification
Sean T. Yang, Bernease Herman, Bill Howe
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 label text embeddings (BERT) as graph node features | Surj · Ontologue COMEDY | AUPRC: 0.857 -> 0.872 | Improved |
| add label text embeddings (BERT) as graph node features | Surj · Ontologue ENGINEERING | AUPRC: 0.853 -> 0.873 | Improved |
| add label text embeddings (BERT) as graph node features | Surj · Ontologue LAW | AUPRC: 0.796 -> 0.802 | Improved |
| add label text embeddings (BERT) as graph node features | Surj · Ontologue MAIN | AUPRC: 0.809 -> 0.834 | Improved |
| add label text embeddings | Surj · Ontologue MAIN | normalized GHV: 0.006 -> 0.005 | Improved |
| add label text embeddings | Surj · Ontologue ENGINEERING | normalized GHV: 0.003 -> 0.002 | Improved |
| use C-HMCNN instead of Surj | Surj · Ontologue ENGINEERING | AUPRC: 0.853 -> 0.842 | Hurt |
| use C-HMCNN instead of Surj | Surj · Ontologue COMEDY | AUPRC: 0.857 -> 0.870 | Improved |
| use Surj instead of Naive baseline for hierarchy compliance | Naive · Ontologue MAIN | normalized GHV: 0.038 -> 0.006 | Improved |
| use CLUS-HMC instead of Naive baseline | Naive · Ontologue COMEDY | AUPRC: 0.839 -> 0.620 | Hurt |
| proposed benchmarks vs prior benchmarks in share of trivial or impossible labels | benchmarks · label difficulty | fraction trivial/impossible labels: 75 % -> 45 % to 58 % | Improved |
Closest papers by meaning
- Learning Semantic-Specific Graph Representation for Multi-Label Image Recognition (ICCV 2019 · similarity 0.86)
- Modeling Label Space Interactions in Multi-label Classification using Box Embeddings (ICLR 2022 · similarity 0.86)
- Knowledge Amalgamation for Multi-Label Classification via Label Dependency Transfer (AAAI 2023 · similarity 0.86)
- End-to-End Ontology Learning with Large Language Models (NeurIPS 2024 · similarity 0.84)
- Distilling Reliable Knowledge for Instance-Dependent Partial Label Learning (AAAI 2024 · similarity 0.83)
- Semantic-Guided Novel Category 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.
