mlsanityData to 7 Oct 2026

Baleen: Robust Multi-Hop Reasoning at Scale via Condensed Retrieval

arXiv 2101.00436 ↗First posted 2 Jan 2021cs.CL cs.IRcited by 63

Not on the map. Why?

Abstract (arXiv)

Multi-hop reasoning (i.e., reasoning across two or more documents) is a key ingredient for NLP models that leverage large corpora to exhibit broad knowledge. To retrieve evidence passages, multi-hop models must contend with a fast-growing search space across the hops, represent complex queries that combine multiple information needs, and resolve ambiguity about the best order in which to hop between training passages. We tackle these problems via Baleen, a system that improves the accuracy of multi-hop retrieval while learning robustly from weak training signals in the many-hop setting. To tame the search space, we propose condensed retrieval, a pipeline that summarizes the retrieved passages after each hop into a single compact context. To model complex queries, we introduce a focused late interaction retriever that allows different parts of the same query representation to match disparate relevant passages. Lastly, to infer the hopping dependencies among unordered training passages, we devise latent hop ordering, a weak-supervision strategy in which the trained retriever itself selects the sequence of hops. We evaluate Baleen on retrieval for two-hop question answering and many-hop claim verification, establishing state-of-the-art performance.

Key facts. Cited by 63 papers in the map (data through 7 Oct 2026). Builds on 26 of its 33 references that are in the map. First posted on arXiv 2 Jan 2021.

Builds on

26 of 33 references.

References · 26

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  4. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,779
  5. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,686
  6. 2017Billion-scale similarity search with GPUscited by 2,477
  7. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,126
  8. 2016MS MARCO: A Human Generated MAchine Reading COmprehension Datasetcited by 1,591
  9. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,362
  10. 2020Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Stepscited by 1,259
  11. 2020ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERTcited by 1,160
  12. 2017Reading Wikipedia to Answer Open-Domain Questionscited by 777
14 more
  1. 2019Latent Retrieval for Weakly Supervised Open Domain Question Answeringcited by 691
  2. 2020KILT: a Benchmark for Knowledge Intensive Language Taskscited by 571
  3. 2018Wizard of Wikipedia: Knowledge-Powered Conversational agentscited by 530
  4. 2020HoVer: A Dataset for Many-Hop Fact Extraction And Claim Verificationcited by 199
  5. 2017Constructing Datasets for Multi-hop Reading Comprehension Across Documentscited by 190
  6. 2020Answering Complex Open-Domain Questions with Multi-Hop Dense Retrievalcited by 160
  7. 2019Learning to Retrieve Reasoning Paths over Wikipedia Graph for Question Answeringcited by 147
  8. 2020Relevance-guided Supervision for OpenQA with ColBERTcited by 71
  9. 2019Answering Complex Open-domain Questions Through Iterative Query Generationcited by 60
  10. 2019Understanding Dataset Design Choices for Multi-hop Reasoningcited by 50
  11. 2019Multi-step Retriever-Reader Interaction for Scalable Open-domain Question Answeringcited by 48
  12. 2019Multi-Hop Paragraph Retrieval for Open-Domain Question Answeringcited by 34
  13. 2020A Memory Efficient Baseline for Open Domain Question Answeringcited by 32
  14. 2019Do Multi-hop Readers Dream of Reasoning Chains?cited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 4; 2022: 10; 2023: 13; 2024: 14; 2025: 10; 2026: 12.

63 papers.

Sort:

By area:

  1. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questionscited by 911
  2. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 803
  3. 2021ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interactioncited by 464
  4. 2022Demonstrate-Search-Predict: Composing retrieval and language models for knowledge-intensive NLPcited by 316
  5. 2024Optimizing Instructions and Demonstrations for Multi-Stage Language Model Programscited by 270
  6. 2024AI Agents That Mattercited by 199
  7. 2022PLAID: An Efficient Engine for Late Interaction Retrievalcited by 95
  8. 2025Recursive Language Modelscited by 94
  9. 2022Iteratively Prompt Pre-trained Language Models for Chain of Thoughtcited by 75
  10. 2024SuRe: Summarizing Retrievals using Answer Candidates for Open-domain QA of LLMscited by 67
  11. 2022Retrieval-Enhanced Machine Learningcited by 46
  12. 2021Joint Passage Ranking for Diverse Multi-Answer Retrievalcited by 43
38 more
  1. 2024Retrieval Augmented Time Series Forecastingcited by 40
  2. 2024Fine-Tuning and Prompt Optimization: Two Great Steps that Work Better Togethercited by 38
  3. 2023UDAPDR: Unsupervised Domain Adaptation via LLM Prompting and Distillation of Rerankerscited by 37
  4. 2021Adaptive Information Seeking for Open-Domain Question Answeringcited by 29
  5. 2023PoET: A generative model of protein families as sequences-of-sequencescited by 29
  6. 2023DSPy Assertions: Computational Constraints for Self-Refining Language Model Pipelinescited by 27
  7. 2024Cofca: A Step-Wise Counterfactual Multi-hop QA benchmarkcited by 26
  8. 2024Claim Verification in the Age of Large Language Models: A Surveycited by 23
  9. 2024StructuredRAG: JSON Response Formatting with Large Language Modelscited by 22
  10. 2022Reasoning over Public and Private Data in Retrieval-Based Systemscited by 21
  11. 2023Chain-of-Skills: A Configurable Model for Open-domain Question Answeringcited by 19
  12. 2024Grounding by Trying: LLMs with Reinforcement Learning-Enhanced Retrievalcited by 17
  13. 2021Cut the CARP: Fishing for zero-shot story evaluationcited by 14
  14. 2023Enhancing LLM Intelligence with ARM-RAG: Auxiliary Rationale Memory for Retrieval Augmented Generationcited by 14
  15. 2022Multi-hop Evidence Retrieval for Cross-document Relation Extractioncited by 13
  16. 2022Claim-Dissector: An Interpretable Fact-Checking System with Joint Re-ranking and Veracity Predictioncited by 12
  17. 2022Natural Logic-guided Autoregressive Multi-hop Document Retrieval for Fact Verificationcited by 11
  18. 2023EXPLAIN, EDIT, GENERATE: Rationale-Sensitive Counterfactual Data Augmentation for Multi-hop Fact Verificationcited by 9
  19. 2025SAGE: A Framework of Precise Retrieval for RAGcited by 8
  20. 2026SoK: Agentic Retrieval-Augmented Generation (RAG): Taxonomy, Architectures, Evaluation, and Research Directionscited by 7
  21. 2022A Retrieve-and-Read Framework for Knowledge Graph Link Predictioncited by 6
  22. 2024HGOT: Hierarchical Graph of Thoughts for Retrieval-Augmented In-Context Learning in Factuality Evaluationcited by 6
  23. 2026SPADER: Step-wise Peer Advantage with Diversity-Aware Exploration Rewards for Multi-Answer Question Answeringcited by 6
  24. 2023Bridging Continuous and Discrete Spaces: Interpretable Sentence Representation Learning via Compositional Operationscited by 5
  25. 2024Language Models Can Reduce Asymmetry in Information Marketscited by 5
  26. 2025MedPAIR: Measuring Physicians and AI Relevance Alignment in Medical Question Answeringcited by 5
  27. 2024Enhancing Cross-Document Event Coreference Resolution by Discourse Structure and Semantic Informationcited by 4
  28. 2025RAAD-LLM: Adaptive Anomaly Detection Using LLMs and RAG Integrationcited by 4
  29. 2026DF-RAG: Query-Aware Diversity for Retrieval-Augmented Generationcited by 4
  30. 2023Consistent Multi-Granular Rationale Extraction for Explainable Multi-hop Fact Verificationcited by 3
  31. 2026CoSearch: Joint Training of Reasoning and Document Ranking via Reinforcement Learning for Agentic Searchcited by 3
  32. 2023LEA: Improving Sentence Similarity Robustness to Typos Using Lexical Attention Biascited by 2
  33. 2023Teaching Smaller Language Models To Generalise To Unseen Compositional Questionscited by 2
  34. 2023Uncertainty Guided Global Memory Improves Multi-Hop Question Answeringcited by 2
  35. 2025GraphCheck: Multipath Fact-Checking with Entity-Relationship Graphscited by 2
  36. 2025PASS-FC: Progressive and Adaptive Search Scheme for Fact Checking of Comprehensive Claimscited by 2
  37. 2025Fact in Fragments: Deconstructing Complex Claims via LLM-based Atomic Fact Extraction and Verificationcited by 2
  38. 2025Cooperative Retrieval-Augmented Generation for Question Answering: Mutual Information Exchange and Ranking by Contrasting Layerscited by 2

50 most cited of 63 All 63 (JSON)