Large Dual Encoders Are Generalizable Retrievers
Jianmo Ni, Chen Qu, Jing Lu, Zhuyun Dai, Gustavo Hernández Ábrego, Ji Ma, Vincent Y. Zhao, Yi Luan, Keith B. Hall, Ming-Wei Chang, Yinfei Yang Google Research Mountain View, CA
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 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, Generalizable T5-based dense Retrievers (GTR), outperform existing sparse and dense retrievers on the BEIR dataset Thakur et al. (2021) significantly. Most surprisingly, our ablation study finds that GTR is very data efficient, as it only needs 10% of MS Marco supervised data to achieve the best out-of-domain performance.111All the GTR models are released at https://tfhub.dev/google/collections/gtr/1.
中文速览
双编码器（dual encoder）检索模型通常被认为因查询向量与文档向量之间只做点积运算，表达能力太弱，难以迁移到训练域之外的任务。研究者挑战这一观点，在保持瓶颈嵌入维度（768维）不变的前提下，将编码器规模扩大至50亿参数（基于T5架构），并结合用20亿条网络问答对做预训练、再用MS Marco精调的两阶段训练策略，构建出名为GTR（Generalizable T5-based dense Retrievers）的系列模型。在涵盖9个领域、18个任务的BEIR零样本泛化基准上，GTR显著超越了现有稀疏检索和密集检索方法，证明单纯扩大模型容量就能突破点积瓶颈的性能上限。更令人惊喜的是，消融实验发现GTR极为数据高效——只需使用10%的MS Marco标注数据，便能达到最优的跨域泛化效果，这对于降低大规模检索系统的标注成本具有重要实践价值。
原文 arXiv:2112.07899;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2112.07899v1