Research map / Architectures and efficiency / Attention and transformers
CopyMTL: Copy Mechanism for Joint Extraction of Entities and Relations with Multi-Task Learning
Daojian Zeng, Haoran Zhang, Qianying Liu
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 linear entity-copying scores with a nonlinear fused-feature layer (ablation) | CopyRE · NYT | triplet F1: .571 -> .722 | Improved |
| replace linear entity-copying scores with a nonlinear fused-feature layer (ablation) | CopyRE · WebNLG | triplet F1: .291 -> .605 | Improved |
| replace linear entity-copying scores with a nonlinear fused-feature layer (ablation) | CopyRE · NYT | entity F1: .647 -> .756 | Improved |
| replace linear entity-copying scores with a nonlinear fused-feature layer (ablation) | CopyRE · WebNLG | entity F1: .595 -> .782 | Improved |
| add a sequence-labeling NER task rather than predict entity length in decoder outputs (ablation) | CopyRE' · NYT | F1: .671 -> .709 | Improved |
| add a sequence-labeling NER task rather than predict entity length in decoder outputs (ablation) | CopyRE' · WebNLG | F1: .553 -> .589 | Improved |
| use separate decoders for each triplet instead of shared decoder parameters (ablation) | CopyMTL · NYT | F1: .709 -> .720 | Improved |
| use separate decoders for each triplet instead of shared decoder parameters (ablation) | CopyMTL · WebNLG | F1: .589 -> .564 | Hurt |
| add the entity mask during inference (ablation) | CopyMTL · NYT and WebNLG | F1 | No clear effect |
| remove the entity mask (ablation) | CopyRE · NYT and WebNLG | F1: zero | Hurt |
| add multi-task sequence labeling to predict complete multi-token entities (ablation) | CopyRE' · NYT and WebNLG | F1: .671 -> .709; .553 -> .589 | Improved |
Closest papers by meaning
- Enhancing Relation Extraction via Supervised Rationale Verification and Feedback (AAAI 2025 · similarity 0.85)
- DocKS-RAG: Optimizing Document-Level Relation Extraction through LLM-Enhanced Hybrid Prompt Tuning (ICML 2025 · similarity 0.85)
- Reliable Data Generation and Selection for Low-Resource Relation Extraction (AAAI 2024 · similarity 0.85)
- A Hierarchical Network for Multimodal Document-Level Relation Extraction (AAAI 2024 · similarity 0.84)
- An Autoregressive Text-to-Graph Framework for Joint Entity and Relation Extraction (AAAI 2024 · similarity 0.84)
- COMM: Concentrated Margin Maximization for Robust Document-Level Relation Extraction (AAAI 2025 · 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.
