# Complementing Lexical Retrieval with Semantic Residual Embedding

arXiv: https://arxiv.org/abs/2004.13969 · HTML: https://mlsanity.com/paper/2004.13969/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Apr 2020.

## Built on by (most cited first, 27 of 27)

- [BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models](https://mlsanity.com/paper/2104.08663/) (2021)
- [Mr. TyDi: A Multi-lingual Benchmark for Dense Retrieval](https://mlsanity.com/paper/2108.08787/) (2021)
- [The Expando-Mono-Duo Design Pattern for Text Ranking with Pretrained Sequence-to-Sequence Models](https://mlsanity.com/paper/2101.05667/) (2021)
- [A Few Brief Notes on DeepImpact, COIL, and a Conceptual Framework for Information Retrieval Techniques](https://mlsanity.com/paper/2106.14807/) (2021)
- [Fine-Grained Distillation for Long Document Retrieval](https://mlsanity.com/paper/2212.10423/) (2022)
- [Extreme Multi-label Learning for Semantic Matching in Product Search](https://mlsanity.com/paper/2106.12657/) (2021)
- [A Proposed Conceptual Framework for a Representational Approach to Information Retrieval](https://mlsanity.com/paper/2110.01529/) (2021)
- [CharacterBERT and Self-Teaching for Improving the Robustness of Dense Retrievers on Queries with Typos](https://mlsanity.com/paper/2204.00716/) (2022)
- [Article Reranking by Memory-Enhanced Key Sentence Matching for Detecting Previously Fact-Checked Claims](https://mlsanity.com/paper/2112.10322/) (2021)
- [Resources for Brewing BEIR: Reproducible Reference Models and an Official Leaderboard](https://mlsanity.com/paper/2306.07471/) (2023)
- [SeDR: Segment Representation Learning for Long Documents Dense Retrieval](https://mlsanity.com/paper/2211.10841/) (2022)
- [Long-tailed Extreme Multi-label Text Classification with Generated Pseudo Label Descriptions](https://mlsanity.com/paper/2204.00958/) (2022)
- [Simple Yet Effective Neural Ranking and Reranking Baselines for Cross-Lingual Information Retrieval](https://mlsanity.com/paper/2304.01019/) (2023)
- [Decentralizing AI Memory: SHIMI, a Semantic Hierarchical Memory Index for Scalable Agent Reasoning](https://mlsanity.com/paper/2504.06135/) (2025)
- [PGT: Pseudo Relevance Feedback Using a Graph-Based Transformer](https://mlsanity.com/paper/2101.07918/) (2021)
- [Divide & Conquer for Entailment-aware Multi-hop Evidence Retrieval](https://mlsanity.com/paper/2311.02616/) (2023)
- [T2-RAGBench: Text-and-Table Benchmark for Evaluating Retrieval-Augmented Generation](https://mlsanity.com/paper/2506.12071/) (2025)
- [Asyncval: A Toolkit for Asynchronously Validating Dense Retriever Checkpoints during Training](https://mlsanity.com/paper/2202.12510/) (2022)
- [Hybrid Inverted Index Is a Robust Accelerator for Dense Retrieval](https://mlsanity.com/paper/2210.05521/) (2022)
- [EncouRAGe: Evaluating RAG Local, Fast, and Reliable](https://mlsanity.com/paper/2511.04696/) (2025)
- [Improving Transformer-Kernel Ranking Model Using Conformer and Query Term Independence](https://mlsanity.com/paper/2104.09393/) (2021)
- [Multifaceted Improvements for Conversational Open-Domain Question Answering](https://mlsanity.com/paper/2204.00266/) (2022)
- [EHI: End-to-end Learning of Hierarchical Index for Efficient Dense Retrieval](https://mlsanity.com/paper/2310.08891/) (2023)
- [Transferring Semantic Knowledge Into Language Encoders](https://mlsanity.com/paper/2110.07382/) (2021)
- [Hybrid and Collaborative Passage Reranking](https://mlsanity.com/paper/2305.09313/) (2023)
- [Multi-word Term Embeddings Improve Lexical Product Retrieval](https://mlsanity.com/paper/2406.01233/) (2024)
- [GliLem: Leveraging GliNER for Contextualized Lemmatization in Estonian](https://mlsanity.com/paper/2412.20597/) (2024)

## Abstract (arXiv)

This paper presents CLEAR, a retrieval model that seeks to complement classical lexical exact-match models such as BM25 with semantic matching signals from a neural embedding matching model. CLEAR explicitly trains the neural embedding to encode language structures and semantics that lexical retrieval fails to capture with a novel residual-based embedding learning method. Empirical evaluations demonstrate the advantages of CLEAR over state-of-the-art retrieval models, and that it can substantially improve the end-to-end accuracy and efficiency of reranking pipelines.