Skip to main content
QUICK REVIEW

[论文解读] Cross Modal Retrieval with Querybank Normalisation

Simion-Vlad Bogolin, Ioana Croitoru|arXiv (Cornell University)|Dec 23, 2021
Multimodal Machine Learning Applications被引用 4
一句话总结

本文提出查询库归一化(Querybank Normalisation, QB-Norm),一种非参数化框架,通过使用查询库中的图像嵌入对查询-图库相似度进行重归一化,缓解了跨模态检索中的中心度问题。该方法引入了动态反向Softmax(Dynamic Inverted Softmax, DIS),实现更鲁棒的相似度归一化,并在无需微调或重新训练的情况下,在多种模型和基准上实现了最先进性能。

ABSTRACT

Profiting from large-scale training datasets, advances in neural architecture design and efficient inference, joint embeddings have become the dominant approach for tackling cross-modal retrieval. In this work we first show that, despite their effectiveness, state-of-the-art joint embeddings suffer significantly from the longstanding "hubness problem" in which a small number of gallery embeddings form the nearest neighbours of many queries. Drawing inspiration from the NLP literature, we formulate a simple but effective framework called Querybank Normalisation (QB-Norm) that re-normalises query similarities to account for hubs in the embedding space. QB-Norm improves retrieval performance without requiring retraining. Differently from prior work, we show that QB-Norm works effectively without concurrent access to any test set queries. Within the QB-Norm framework, we also propose a novel similarity normalisation method, the Dynamic Inverted Softmax, that is significantly more robust than existing approaches. We showcase QB-Norm across a range of cross modal retrieval models and benchmarks where it consistently enhances strong baselines beyond the state of the art. Code is available at https://vladbogo.github.io/QB-Norm/.

研究动机与目标

  • 解决在用于跨模态检索的高维联合嵌入中持续存在的中心度问题。
  • 开发一种无需模型微调或重新训练即可提升性能的检索框架。
  • 在推理阶段实现有效的中心度缓解,而无需同时访问多个测试查询。
  • 设计一种对查询库选择不敏感的鲁棒相似度归一化方法。
  • 在多种模型、模态和基准上实现泛化能力。

提出的方法

  • QB-Norm 使用来自图库嵌入的查询库,在推理阶段对查询与图库样本之间的相似度分数进行重归一化,以降低中心节点的影响。
  • 查询库在推理期间从图库嵌入中构建,实现无需重新训练的中心度校正。
  • 引入动态反向Softmax(DIS)作为新颖的归一化机制,根据中心频率自适应调整相似度分数。
  • DIS 使用反温度参数控制软化的程度,β=20时性能最优。
  • 该框架在推理后应用,因此可与任何预训练的跨模态嵌入模型兼容。
  • QB-Norm 仅基于相似度分数运行,无需架构修改或额外训练。
Figure 1 : Left: The hubness problem. We consider the problem of cross modal retrieval in which queries $q_{1}$ and $q_{2}$ are compared against a gallery of samples, $x_{1}$ and $x_{2}$ . As we show in Sec. 3.2 , the high-dimensional joint embeddings employed by modern methods for cross-modal retri
Figure 1 : Left: The hubness problem. We consider the problem of cross modal retrieval in which queries $q_{1}$ and $q_{2}$ are compared against a gallery of samples, $x_{1}$ and $x_{2}$ . As we show in Sec. 3.2 , the high-dimensional joint embeddings employed by modern methods for cross-modal retri

实验结果

研究问题

  • RQ1在现代跨模态检索系统中,是否可以在不重新训练的情况下有效缓解中心度问题?
  • RQ2当推理时仅有一个测试查询可用时,QB-Norm 是否依然有效?
  • RQ3能否设计一种对查询库选择鲁棒的相似度归一化方法,并在多种模型上提升检索性能?
  • RQ4QB-Norm 与现有中心度缓解技术相比,在性能和鲁棒性方面表现如何?
  • RQ5QB-Norm 是否能在不同模态和检索基准上实现泛化?

主要发现

  • 在MSR-VTT v2t检索任务中,QB-Norm将TT-CE+模型的R@1绝对提升5.9%,从24.6%提高到30.1%。
  • 在同一基准上,R@5从54.1%提升至61.4%,R@10从67.5%提升至73.2%。
  • 该方法在文本到视频、文本到图像以及音频到文本检索任务中均一致提升了检索性能。
  • 随着查询库增大,性能趋于饱和,表明超过一定规模后收益递减。
  • 动态反向Softmax(DIS)优于先前的归一化方法,且对查询库选择具有鲁棒性,避免了在某些查询库上性能下降。
  • 即使推理时仅有一个查询可用,QB-Norm 仍保持有效性,而以往方法需依赖多个查询。
Figure 2 : Hubness is pervasive in leading methods for text-video retrieval . The charts depict the distribution of the number of times each gallery video was retrieved by test set queries (x-axes video ids are ordered by decreasing retrieval count). Top row (different models): We report retrieval d
Figure 2 : Hubness is pervasive in leading methods for text-video retrieval . The charts depict the distribution of the number of times each gallery video was retrieved by test set queries (x-axes video ids are ordered by decreasing retrieval count). Top row (different models): We report retrieval d

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。