Research map / Graph learning / Knowledge and temporal graphs
Encoder-Decoder Based Unified Semantic Role Labeling with Label-Aware Syntax
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use the unified pointer-network SRL framework with LA-GCN | CoNLL09 English unified SRL · CoNLL09 English WSJ and Brown | argument role F1: 92.2 -> 92.5; 85.2 -> 85.4 | Helped |
| use the unified pointer-network SRL framework | UPB unified SRL · UPB, eight-language average | F1: 83.6 -> 84.9 | Helped |
| use the unified pointer-network SRL framework | UPB unified SRL · UPB Italian | F1: 81.3 -> 80.7 | Hurt |
| add POS tag embeddings (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 92.6 -> 93.2 | Helped |
| add BERT representations (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 90.8 -> 93.2 | Helped |
| add LA-GCN syntax encoding (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 91.2 -> 93.2 | Helped |
| encode dependency labels as well as arcs with LA-GCN instead of GCN* (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 91.9 -> 93.2 | Helped |
| use LA-GCN instead of LP-GCN (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 92.7 -> 93.2 | Helped |
| add the syntactic-aware pointer (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 92.2 -> 93.2 | Helped |
| add high-order interacted attention (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 92.0 -> 93.2 | Helped |
| use the enhanced pointer instead of the base pointer (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 91.7 -> 93.2 | Helped |
| use close-first rather than left-to-right decoding order (ablation) | Ours on CoNLL09 English WSJ development · CoNLL09 English WSJ development | F1: 92.5 -> 93.2 | Helped |
Closest papers by meaning
- End-to-end Semantic Role Labeling with Neural Transition-based Model (AAAI 2021 · similarity 0.91)
- Syntax-Aware Neural Semantic Role Labeling (AAAI 2019 · similarity 0.91)
- Deep Semantic Role Labeling With Self-Attention (AAAI 2018 · similarity 0.90)
- Span-Based Semantic Role Labeling with Argument Pruning and Second-Order Inference (AAAI 2022 · similarity 0.90)
- Frame Semantic Role Labeling Using Arbitrary-Order Conditional Random Fields (AAAI 2024 · similarity 0.88)
- UniGTE: Unified Graph–Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains (NeurIPS 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.