Research map / Large language models / Retrieval and knowledge
Multi-step Retriever-Reader Interaction for Scalable Open-domain Question Answering
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 multi-step retriever-reader interaction | DrQA · Quasar-T | EM: 36.87 -> 39.53 | Helped |
| Add multi-step retriever-reader interaction | DrQA · SearchQA | F1: 58.24 -> 61.61 | Helped |
| Add multi-step retriever-reader interaction | DrQA · TriviaQA-unfiltered | F1: 52.13 -> 61.66 | Helped |
| Add multi-step retriever-reader interaction | DrQA · SQuAD-open | EM: 27.1 -> 31.93 | Helped |
| Add 3 steps of retriever-reader interaction | DrQA · TriviaQA-open | F1: 42.16 -> 44.30 | Helped |
| Vary the number of retriever-reader interaction steps from none to multiple steps (ablation) | DrQA multi-step reasoner · Quasar-T, SearchQA, TriviaQA-unfiltered | F1 | Helped |
| Increase interaction steps beyond the best-performing range (ablation) | DrQA multi-step reasoner · Quasar-T, SearchQA, TriviaQA-unfiltered | F1 | No clear effect |
| Add a minibatch-average reward baseline to REINFORCE training (ablation) | Multi-step reasoner · Open-domain QA | QA performance | Hurt |
| Pre-train the multi-step reasoner before reinforcement learning fine-tuning (ablation) | Multi-step reasoner · Open-domain QA | QA performance | Helped |
Closest papers by meaning
- Retriever-Free Retrieval-Augmented Reasoning via Corpus-Traversing MCTS (NeurIPS 2026 · similarity 0.87)
- SuRe: Summarizing Retrievals using Answer Candidates for Open-domain QA of LLMs (ICLR 2024 · similarity 0.86)
- RVR: Retrieve-Verify-Retrieve for Comprehensive Question Answering (NeurIPS 2026 · similarity 0.85)
- BRIGHT: A Realistic and Challenging Benchmark for Reasoning-Intensive Retrieval (ICLR 2025 · similarity 0.85)
- Cooperative Retrieval-Augmented Generation for Question Answering: Mutual Information Exchange and Ranking by Contrasting Layers (NeurIPS 2025 · similarity 0.85)
- Autofocus Retrieval: An Effective Pipeline for Multi-Hop Question Answering With Semi-Structured Knowledge (NeurIPS 2026 · similarity 0.85)
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.