Research map / Graph learning / Knowledge and temporal graphs
Task-Specific Representation Learning for Web-Scale Entity Disambiguation
Rijula Kar, Susmija Reddy, Sourangshu Bhattacharya, Anirban Dasgupta, Soumen Chakrabarti
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 entity definition page text features (ablation) | TSRL · AIDA-CoNLL testb | micro-avg accuracy: 88.0 -> 89.4 | Improved |
| add entity definition page text features (ablation) | TSRL · AIDA-CoNLL testb | macro-avg accuracy: 90.2 -> 91.9 | Improved |
| task-sensitive representation (TSRL) instead of per-alias classifier (ablation) | PAC · AIDA-CoNLL testb | micro-avg accuracy: 85.9 -> 88.0 | Improved |
| task-sensitive representation instead of signhash MTL (ablation) | MTL-SH · AIDA-CoNLL testb | micro-avg accuracy: 86.9 -> 88.0 | Improved |
| task-sensitive representation instead of task-independent MTRL (ablation) | MTRL · AIDA-CoNLL testb | micro-avg accuracy: 82.0 -> 88.0 | Improved |
| multitask sharing across mentions via signhash (ablation) | PAC · AIDA-CoNLL testb | micro-avg accuracy: 85.9 -> 86.9 | Improved |
| TSRL on gold mentions vs Ganea and Hofmann 2017 (ablation) | TSRL · AIDA-CoNLL testb | micro-avg accuracy: 88.8 -> 88.0 | Hurt |
| TSRL vs PAC on top-prior entities (ablation) | PAC · CoNLL Bin-1 | accuracy: 93.9 % -> 95.5% | Improved |
| TSRL vs PAC on rarer entities (ablation) | PAC · CoNLL Bin-2 | accuracy: 62.4 % -> 68.53% | Improved |
| TSRL vs PAC on entities ranked 11-20 (ablation) | PAC · CoNLL Bin-3 | accuracy: 63.4% -> 73% | Improved |
| TSRL vs PAC on entities ranked 21-30 (ablation) | PAC · CoNLL Bin-4 | accuracy: 64 % -> 64% | No clear effect |
| TSRL vs best prior system (ablation) | TSRL · MSNBC | micro F1: 93.7 -> 94.2 | Improved |
Closest papers by meaning
- A Recurrent Model for Collective Entity Linking with Adaptive Features (AAAI 2020 · similarity 0.90)
- Domain-Adapted Dependency Parsing for Cross-Domain Named Entity Recognition (AAAI 2023 · similarity 0.88)
- A Hierarchical Multi-Task Approach for Learning Embeddings from Semantic Tasks (AAAI 2019 · similarity 0.87)
- Select and Augment: Enhanced Dense Retrieval Knowledge Graph Augmentation (Abstract Reprint) (AAAI 2024 · similarity 0.86)
- UniversalNER: Targeted Distillation from Large Language Models for Open Named Entity Recognition (ICLR 2024 · similarity 0.85)
- Span Graph Transformer for Document-Level Named Entity Recognition (AAAI 2024 · similarity 0.85)
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.
