mlsanityData to 7 Oct 2026

Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval

arXiv 2104.05883 ↗First posted 13 Apr 2021cs.CLcited by 25

Not on the map. Why?

Abstract (arXiv)

Complex question answering often requires finding a reasoning chain that consists of multiple evidence pieces. Current approaches incorporate the strengths of structured knowledge and unstructured text, assuming text corpora is semi-structured. Building on dense retrieval methods, we propose a new multi-step retrieval approach (BeamDR) that iteratively forms an evidence chain through beam search in dense representations. When evaluated on multi-hop question answering, BeamDR is competitive to state-of-the-art systems, without using any semi-structured information. Through query composition in dense space, BeamDR captures the implicit relationships between evidence in the reasoning chain. The code is available at https://github.com/ henryzhao5852/BeamDR.

Key facts. Cited by 25 papers in the map (data through 7 Oct 2026). Builds on 19 of its 32 references that are in the map. First posted on arXiv 13 Apr 2021.

Builds on

19 of 32 references.

References · 19

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,779
  3. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,686
  4. 2017Billion-scale similarity search with GPUscited by 2,477
  5. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,126
  6. 2020Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrievalcited by 846
  7. 2019Latent Retrieval for Weakly Supervised Open Domain Question Answeringcited by 691
  8. 2019Passage Re-ranking with BERTcited by 629
  9. 2018TVQA: Localized, Compositional Video Question Answeringcited by 463
  10. 2018The Web as a Knowledge-base for Answering Complex Questionscited by 409
  11. 2020AmbigQA: Answering Ambiguous Open-domain Questionscited by 385
  12. 2017Variational Reasoning for Question Answering with Knowledge Graphcited by 170
7 more
  1. 2020Answering Complex Open-Domain Questions with Multi-Hop Dense Retrievalcited by 160
  2. 2019Learning to Retrieve Reasoning Paths over Wikipedia Graph for Question Answeringcited by 147
  3. 2014Asymmetric LSH (ALSH) for Sublinear Time Maximum Inner Product Search (MIPS)cited by 109
  4. 2019Answering Complex Open-domain Questions Through Iterative Query Generationcited by 60
  5. 2019Revealing the Importance of Semantic Retrieval for Machine Reading at Scalecited by 50
  6. 2020Differentiable Reasoning over a Virtual Knowledge Basecited by 39
  7. 2021Complex Factoid Question Answering with a Free-Text Knowledge Graphcited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 8; 2022: 7; 2023: 3; 2024: 3; 2025: 2; 2026: 2.

25 papers.

Sort:

By area:

  1. 2021A Dataset for Answering Time-Sensitive Questionscited by 150
  2. 2022Generating Natural Language Proofs with Verifier-Guided Searchcited by 52
  3. 2021Mention Memory: incorporating textual knowledge into Transformers through entity mention attentioncited by 42
  4. 2023End-to-End Beam Retrieval for Multi-Hop Question Answeringcited by 35
  5. 2022Entailment Tree Explanations via Iterative Retrieval-Generation Reasonercited by 26
  6. 2021What's in a Name? Answer Equivalence For Open-Domain Question Answeringcited by 21
  7. 2022Few-shot Reranking for Multi-hop QA via Language Model Promptingcited by 19
  8. 2021Hybrid Autoregressive Inference for Scalable Multi-hop Explanation Regenerationcited by 17
  9. 2021UNIQORN: Unified Question Answering over RDF Knowledge Graphs and Natural Language Textcited by 17
  10. 2022Open-domain Question Answering via Chain of Reasoning over Heterogeneous Knowledgecited by 14
  11. 2024Reverse Question Answering: Can an LLM Write a Question so Hard (or Bad) that it Can't Answer?cited by 13
  12. 2025KiRAG: Knowledge-Driven Iterative Retriever for Enhancing Retrieval-Augmented Generationcited by 12
13 more
  1. 2021Distantly-Supervised Evidence Retrieval Enables Question Answering without Evidence Annotationcited by 5
  2. 2023How Many Answers Should I Give? An Empirical Study of Multi-Answer Reading Comprehensioncited by 5
  3. 2022ReasonChainQA: Text-based Complex Question Answering with Explainable Evidence Chainscited by 4
  4. 2021Extract, Integrate, Compete: Towards Verification Style Reading Comprehensioncited by 3
  5. 2022Better Retrieval May Not Lead to Better Question Answeringcited by 3
  6. 2021Interactive Machine Comprehension with Dynamic Knowledge Graphscited by 2
  7. 2023Reimagining Retrieval Augmented Language Models for Answering Queriescited by 2
  8. 2025Gumbel Reranking: Differentiable End-to-End Reranker Optimizationcited by 2
  9. 2026CIRAG: Construction-Integration Retrieval and Adaptive Generation for Multi-hop Question Answeringcited by 2
  10. 2024Large Language Models Still Face Challenges in Multi-Hop Reasoning with External Knowledgecited by 1
  11. 2024Momentum Posterior Regularization for Multi-hop Dense Retrievalcited by 1
  12. 2022Cheater's Bowl: Human vs. Computer Search Strategies for Open-Domain Question Answeringcited by 0
  13. 2026Forgotten History or Test-of-Time? Retrospect and Prospect on RAG from an IR Perspectivecited by 0