Retrieve Fast, Rerank Smart: Cooperative and Joint Approaches for Improved Cross-Modal Retrieval

• Published in: Transactions of the Association for Computational Linguistics Vol. 10; pp. 503 - 521
• Main Authors: Geigle, Gregor, Pfeiffer, Jonas, Reimers, Nils, Vulić, Ivan, Gurevych, Iryna
• Format: Journal Article
• Language: English
• Published: One Broadway, 12th Floor, Cambridge, Massachusetts 02142, USA MIT Press 04.05.2022; MIT Press Journals, The; The MIT Press
• Subjects: Benchmarks; Coders; Computational linguistics; Efficiency; Image retrieval; Linguistics; Multilingualism
• ISSN: 2307-387X; 2307-387X
• DOI: 10.1162/tacl_a_00473

Current state-of-the-art approaches to cross- modal retrieval process text and visual input jointly, relying on Transformer-based architectures with cross-attention mechanisms that attend over all words and objects in an image. While offering unmatched retrieval performance, such models: are typically pretrained from scratch and thus less scalable, suffer from huge retrieval latency and inefficiency issues, which makes them impractical in realistic applications. To address these crucial gaps towards both improved and efficient cross- modal retrieval, we propose a novel fine-tuning framework that turns any pretrained text-image multi-modal model into an efficient retrieval model. The framework is based on a cooperative retrieve-and-rerank approach that combines: twin networks (i.e., a bi-encoder) to separately encode all items of a corpus, enabling efficient initial retrieval, and a cross-encoder component for a more nuanced (i.e., smarter) ranking of the retrieved small set of items. We also propose to jointly fine- tune the two components with shared weights, yielding a more parameter-efficient model. Our experiments on a series of standard cross-modal retrieval benchmarks in monolingual, multilingual, and zero-shot setups, demonstrate improved accuracy and huge efficiency benefits over the state-of-the-art cross- encoders.