LightningDOT: Pre-training Visual-Semantic Embeddings for Real-Time Image-Text Retrieval

Siqi Sun, Yen-Chun Chen, Linjie Li, Shuohang Wang, Yuwei Fang, Jingjing Liu


Abstract
Multimodal pre-training has propelled great advancement in vision-and-language research. These large-scale pre-trained models, although successful, fatefully suffer from slow inference speed due to enormous computational cost mainly from cross-modal attention in Transformer architecture. When applied to real-life applications, such latency and computation demand severely deter the practical use of pre-trained models. In this paper, we study Image-text retrieval (ITR), the most mature scenario of V+L application, which has been widely studied even prior to the emergence of recent pre-trained models. We propose a simple yet highly effective approach, LightningDOT that accelerates the inference time of ITR by thousands of times, without sacrificing accuracy. LightningDOT removes the time-consuming cross-modal attention by extracting pre-cached feature indexes offline, and employing instant dot-product matching online, which significantly speeds up retrieval process. In fact, our LightningDOT achieves superior performance across mainstream ITR benchmarks such as Flickr30k and COCO datasets, outperforming existing pre-trained models that consume 1000 times magnitude of computational hours using the same features.
Anthology ID:
2021.naacl-main.77
Volume:
Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Month:
June
Year:
2021
Address:
Online
Editors:
Kristina Toutanova, Anna Rumshisky, Luke Zettlemoyer, Dilek Hakkani-Tur, Iz Beltagy, Steven Bethard, Ryan Cotterell, Tanmoy Chakraborty, Yichao Zhou
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
982–997
Language:
URL:
https://aclanthology.org/2021.naacl-main.77
DOI:
10.18653/v1/2021.naacl-main.77
Bibkey:
Cite (ACL):
Siqi Sun, Yen-Chun Chen, Linjie Li, Shuohang Wang, Yuwei Fang, and Jingjing Liu. 2021. LightningDOT: Pre-training Visual-Semantic Embeddings for Real-Time Image-Text Retrieval. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 982–997, Online. Association for Computational Linguistics.
Cite (Informal):
LightningDOT: Pre-training Visual-Semantic Embeddings for Real-Time Image-Text Retrieval (Sun et al., NAACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.naacl-main.77.pdf
Video:
 https://aclanthology.org/2021.naacl-main.77.mp4
Code
 intersun/LightningDOT +  additional community code
Data
MS COCO