Research map / Large language models / Retrieval and knowledge
A Deep Sequential Model for Discourse Parsing on Multi-Party Dialogues
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 |
|---|---|---|---|
| Jointly predict links and relations while incrementally encoding the predicted structure | Deep Sequential versus strongest baseline Deep+ILP · STAC test | Link & Rel F1 (%): 53.1 -> 55.7 | Helped |
| Use separate discourse representations for link prediction and relation classification instead of sharing them (ablation) | Deep Sequential · STAC test | Link & Rel F1 (%): 54.7 -> 55.7 | Helped |
| Add structured representations encoding the predicted discourse structure (ablation) | Deep Sequential · STAC test | Link & Rel F1 (%): 53.7 -> 55.7 | Helped |
| Add structured representations encoding the predicted discourse structure (ablation) | Deep Sequential · STAC test | Link F1 (%): 71.0 -> 73.2 | Helped |
| Replace the predicted discourse structure with a random structure (ablation) | Deep Sequential · STAC test | Link F1 (%): 71.8 -> 73.2 | Hurt |
| Replace the predicted discourse structure with a random structure (ablation) | Deep Sequential · STAC test | Link & Rel F1 (%): 53.7 -> 55.7 | Hurt |
| Add random structured representations instead of omitting structured representations (ablation) | Deep Sequential · STAC test | Link F1 (%): 71.0 -> 71.8 | Helped |
| Add the speaker highlighting mechanism (ablation) | Deep Sequential · STAC test | Link & Rel F1 (%): 54.5 -> 55.7 | Helped |
| Add the speaker highlighting mechanism (ablation) | Deep Sequential · STAC test | Link F1 (%): 71.7 -> 73.2 | Helped |
Closest papers by meaning
- Unsupervised Learning of Discourse Structures using a Tree Autoencoder (AAAI 2021 · similarity 0.89)
- Linguistic Properties Matter for Implicit Discourse Relation Recognition: Combining Semantic Interaction, Topic Continuity and Attribution (AAAI 2018 · similarity 0.87)
- A Label Dependence-Aware Sequence Generation Model for Multi-Level Implicit Discourse Relation Recognition (AAAI 2022 · similarity 0.87)
- Predicting Above-Sentence Discourse Structure Using Distant Supervision from Topic Segmentation (AAAI 2022 · similarity 0.87)
- Improving Implicit Discourse Relation Recognition with Natural Language Explanations from LLMs (AAAI 2026 · similarity 0.86)
- Harnessing Holistic Discourse Features and Triadic Interaction for Sentiment Quadruple Extraction in Dialogues (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.