Large Dual Encoders Are Generalizable Retrievers

Abstract

It has been shown that dual encoders trained on one domain often fail to generalize to other domains for retrieval tasks. One widespread belief is that the bottleneck layer of a dual encoder, where the final score is simply a dot-product between a query vector and a passage vector, is too limited to make dual encoders an effective retrieval model for out-of-domain generalization. In this paper, we challenge this belief by scaling up the size of the dual encoder model \{\em while keeping the bottleneck embedding size fixed.\} With multi-stage training, surprisingly, scaling up the model size brings significant improvement on a variety of retrieval tasks, especially for out-of-domain generalization. Experimental results show that our dual encoders, \textbf\{G\}eneralizable \textbf\{T\}5-based dense \textbf\{R\}etrievers (GTR), outperform %ColBERT~\cite\{khattab2020colbert\} and existing sparse and dense retrievers on the BEIR dataset~\cite\{thakur2021beir\} significantly. Most surprising

Stats

• citations90
• S2 citations—
• github stars0
• HF likes0
• heat score14.69
• arxiv keyni2021large

Related papers

• Back To Basics: A Simple Recipe For Improving Out-of-domain Retrieval In Dense Encoders (2023)0.00
• Generalization Properties Of Retrieval-based Models (2022)0.00
• Query Encoder Distillation Via Embedding Alignment Is A Strong Baseline Method To Boost Dense Retriever Online Efficiency (2023)0.00
• How To Train Your DRAGON: Diverse Augmentation Towards Generalizable Dense Retrieval (2023)11.39
• BERM: Training The Balanced And Extractable Representation For Matching To Improve Generalization Ability Of Dense Retrieval (2023)5.84
• Dense Retrievers Can Fail On Simple Queries: Revealing The Granularity Dilemma Of Embeddings (2025)2.86
• Empowering Dual-encoder With Query Generator For Cross-lingual Dense Retrieval (2023)6.34
• Investigating Multi-layer Representations For Dense Passage Retrieval (2025)0.00