Sparse and Dense Retrievers Learn Better Together: Joint Sparse-Dense Optimization for Text-Image Retrieval

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Song, Jonghyun, Lee, Youngjune, Cho, Gyu-Hwung, Song, Ilhyeon, Kim, Saehun, Jo, Yohan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915457748959232
author Song, Jonghyun
Lee, Youngjune
Cho, Gyu-Hwung
Song, Ilhyeon
Kim, Saehun
Jo, Yohan
author_facet Song, Jonghyun
Lee, Youngjune
Cho, Gyu-Hwung
Song, Ilhyeon
Kim, Saehun
Jo, Yohan
contents Vision-Language Pretrained (VLP) models have achieved impressive performance on multimodal tasks, including text-image retrieval, based on dense representations. Meanwhile, Learned Sparse Retrieval (LSR) has gained traction in text-only settings due to its interpretability and efficiency with fast term-based lookup via inverted indexes. Inspired by these advantages, recent work has extended LSR to the multimodal domain. However, these methods often rely on computationally expensive contrastive pre-training, or distillation from a frozen dense model, which limits the potential for mutual enhancement. To address these limitations, we propose a simple yet effective framework that enables bi-directional learning between dense and sparse representations through Self-Knowledge Distillation. This bi-directional learning is achieved using an integrated similarity score-a weighted sum of dense and sparse similarities-which serves as a shared teacher signal for both representations. To ensure efficiency, we fine-tune the final layer of the dense encoder and the sparse projection head, enabling easy adaptation of any existing VLP model. Experiments on MSCOCO and Flickr30k demonstrate that our sparse retriever not only outperforms existing sparse baselines, but also achieves performance comparable to-or even surpassing-its dense counterparts, while retaining the benefits of sparse models.
format Preprint
id arxiv_https___arxiv_org_abs_2508_16707
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sparse and Dense Retrievers Learn Better Together: Joint Sparse-Dense Optimization for Text-Image Retrieval
Song, Jonghyun
Lee, Youngjune
Cho, Gyu-Hwung
Song, Ilhyeon
Kim, Saehun
Jo, Yohan
Computation and Language
Information Retrieval
Machine Learning
Vision-Language Pretrained (VLP) models have achieved impressive performance on multimodal tasks, including text-image retrieval, based on dense representations. Meanwhile, Learned Sparse Retrieval (LSR) has gained traction in text-only settings due to its interpretability and efficiency with fast term-based lookup via inverted indexes. Inspired by these advantages, recent work has extended LSR to the multimodal domain. However, these methods often rely on computationally expensive contrastive pre-training, or distillation from a frozen dense model, which limits the potential for mutual enhancement. To address these limitations, we propose a simple yet effective framework that enables bi-directional learning between dense and sparse representations through Self-Knowledge Distillation. This bi-directional learning is achieved using an integrated similarity score-a weighted sum of dense and sparse similarities-which serves as a shared teacher signal for both representations. To ensure efficiency, we fine-tune the final layer of the dense encoder and the sparse projection head, enabling easy adaptation of any existing VLP model. Experiments on MSCOCO and Flickr30k demonstrate that our sparse retriever not only outperforms existing sparse baselines, but also achieves performance comparable to-or even surpassing-its dense counterparts, while retaining the benefits of sparse models.
title Sparse and Dense Retrievers Learn Better Together: Joint Sparse-Dense Optimization for Text-Image Retrieval
topic Computation and Language
Information Retrieval
Machine Learning
url https://arxiv.org/abs/2508.16707