English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this work, we present a post-processing solution to address the hubnessproblem in cross-modal retrieval, a phenomenon where a small number of gallerydata points are frequently retrieved, resulting in a decline in retrievalperformance. We first theoretically demonstrate the necessity of incorporatingboth the gallery and query data for addressing hubness as hubs always exhibithigh similarity with gallery and query data. Second, building on ourtheoretical results, we propose a novel framework, Dual Bank Normalization(DBNorm). While previous work has attempted to alleviate hubness by onlyutilizing the query samples, DBNorm leverages two banks constructed from thequery and gallery samples to reduce the occurrence of hubs during inference.Next, to complement DBNorm, we introduce two novel methods, dual invertedsoftmax and dual dynamic inverted softmax, for normalizing similarity based onthe two banks. Specifically, our proposed methods reduce the similarity betweenhubs and queries while improving the similarity between non-hubs and queries.Finally, we present extensive experimental results on diverse language-groundedbenchmarks, including text-image, text-video, and text-audio, demonstrating thesuperior performance of our approaches compared to previous methods inaddressing hubness and boosting retrieval performance. Our code is available athttps://github.com/yimuwangcs/Better_Cross_Modal_Retrieval.

Balance Act: Mitigating Hubness in Cross-Modal Retrieval with Query and Gallery Banks