Research map / Label-efficient and robust learning / Recognition and long-tail
End-to-end Semantic Role Labeling with Neural Transition-based Model
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 |
|---|---|---|---|
| Add incremental high-order argument-predicate scoring features (ablation) | Transition SRL · UPB Avg. | F1: 72.9 -> 74.2 | Helped |
| Add dependency-tree features (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 89.8 -> 91.0 | Helped |
| Add POS features (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 90.4 -> 91.0 | Helped |
| Add character encoder (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 90.7 -> 91.0 | Helped |
| Add high-order action feature o^a (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 89.4 -> 91.0 | Helped |
| Add high-order role feature o^r (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 88.8 -> 91.0 | Helped |
| Use close-first parsing instead of left-to-right parsing (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 88.5 -> 91.0 | Helped |
| Use close-first parsing instead of right-to-left parsing (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 88.0 -> 91.0 | Helped |
| Use BERT contextualized word representations | High-order transition SRL · CoNLL09 with contextualized representations | F1 | Helped |
| Use XLNet contextualized word representations | High-order transition SRL · CoNLL09 with contextualized representations | F1 | Helped |
| Use incremental high-order interaction rather than the vanilla transition model (ablation) | Transition SRL · CoNLL09 development set | Arg. F1: 88.6 -> 91.0 | Helped |
Closest papers by meaning
- Encoder-Decoder Based Unified Semantic Role Labeling with Label-Aware Syntax (AAAI 2021 · similarity 0.91)
- Dependency or Span, End-to-End Uniform Semantic Role Labeling (AAAI 2019 · similarity 0.91)
- Deep Semantic Role Labeling With Self-Attention (AAAI 2018 · similarity 0.89)
- Span-Based Semantic Role Labeling with Argument Pruning and Second-Order Inference (AAAI 2022 · similarity 0.89)
- Syntax-Aware Neural Semantic Role Labeling (AAAI 2019 · similarity 0.89)
- Frame Semantic Role Labeling Using Arbitrary-Order Conditional Random Fields (AAAI 2024 · similarity 0.88)
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.