Research map / Large language models / Retrieval and knowledge
Dialog Inpainting: Turning Documents into Dialogs
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 |
|---|---|---|---|
| Pre-train the retriever on WikiDialog | T5-Large dual encoder · QReCC | MRR: 55.7 -> 60.4 | Helped |
| Pre-train the retriever on WikiDialog and WebDialog rather than WikiDialog alone (ablation) | T5-Large dual encoder · QReCC | MRR: 60.4 -> 60.7 | Helped |
| Pre-train the retriever on WikiDialog | T5-Large dual encoder · OR-QuAC | MRR@5: 56.9 -> 66.5 | Helped |
| Pre-train the retriever on WikiDialog and WebDialog rather than WikiDialog alone (ablation) | T5-Large dual encoder · OR-QuAC | MRR@5: 66.5 -> 68.7 | Helped |
| Pre-train the retriever on WikiDialog | T5-Large dual encoder · TREC CAsT-19 | MRR: 61.0 -> 68.1 | Helped |
| Pre-train the retriever on WikiDialog and WebDialog rather than WikiDialog alone (ablation) | T5-Large dual encoder · TREC CAsT-19 | MRR: 68.1 -> 74.1 | Helped |
| Pre-train the retriever on WikiDialog | T5-Large dual encoder · TREC CAsT-20 | MRR: 34.3 -> 43.7 | Helped |
| Pre-train the retriever on WikiDialog and WebDialog rather than WikiDialog alone (ablation) | T5-Large dual encoder · TREC CAsT-20 | MRR: 43.7 -> 51.3 | Helped |
| Pre-train the retriever and reranker on inpainted WikiDialog | T5-Large retrieval and reranking system · OR-QuAC | MRR@5: 72.6 -> 79.7 | Helped |
| Pre-train the retriever and reranker on inpainted WikiDialog | T5-Large retrieval and reranking system · TREC CAsT-19 | MRR: 75.3 -> 79.3 | Helped |
| Pre-train the retriever and reranker on WikiDialog and WebDialog rather than WikiDialog alone (ablation) | T5-Large retrieval and reranking system · TREC CAsT-20 | MRR: 60.3 -> 59.7 | Hurt |
| Generate WikiDialog with an inpainter trained on ConvQA data rather than open-domain dialog data only (ablation) | T5-Base dual encoder · OR-QuAC | MRR@5: 64.0 -> 65.3 | Helped |
Closest papers by meaning
- Complex Sequential Question Answering: Towards Learning to Converse Over Linked Question Answer Pairs with a Knowledge Graph (AAAI 2018 · similarity 0.89)
- DialogLM: Pre-trained Model for Long Dialogue Understanding and Summarization (AAAI 2022 · similarity 0.88)
- Dialog-to-Action: Conversational Question Answering Over a Large-Scale Knowledge Base (NeurIPS 2018 · similarity 0.88)
- Reasoning in Dialog: Improving Response Generation by Context Reading Comprehension (AAAI 2021 · similarity 0.87)
- ChatterBox: Multimodal Referring and Grounding with Chain-of-Questions (AAAI 2025 · similarity 0.86)
- A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering (NeurIPS 2026 · similarity 0.86)
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.