Bi-Encoder Cascades for Efficient Image Search

Robert Hönig, Jan Ackermann, Mingyuan Chi

ICCVW 2023 pp. 1350-1355

doi:10.1109/ICCVW60793.2023.00146 /iccvw/2023/honig2023iccvw-biencoder/

Abstract

Modern neural encoders offer unprecedented text-image retrieval (TIR) accuracy, but their high computational cost impedes an adoption to large-scale image searches. To lower this cost, model cascades use an expensive encoder to refine the ranking of a cheap encoder. However, existing cascading algorithms focus on cross-encoders, which jointly process text-image pairs, but do not consider cascades of bi-encoders, which separately process texts and images. We introduce the small-world search scenario as a realistic setting where bi-encoder cascades can reduce costs. We then propose a cascading algorithm that leverages the small-world search scenario to reduce lifetime image encoding costs of a TIR system. Our experiments show cost reductions by up to 6x.

PDF ICCVW Semantic Scholar

Cite

Text

Hönig et al. "Bi-Encoder Cascades for Efficient Image Search." IEEE/CVF International Conference on Computer Vision Workshops, 2023. doi:10.1109/ICCVW60793.2023.00146

Markdown

[Hönig et al. "Bi-Encoder Cascades for Efficient Image Search." IEEE/CVF International Conference on Computer Vision Workshops, 2023.](https://mlanthology.org/iccvw/2023/honig2023iccvw-biencoder/) doi:10.1109/ICCVW60793.2023.00146

BibTeX

@inproceedings{honig2023iccvw-biencoder,
  title     = {{Bi-Encoder Cascades for Efficient Image Search}},
  author    = {Hönig, Robert and Ackermann, Jan and Chi, Mingyuan},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2023},
  pages     = {1350-1355},
  doi       = {10.1109/ICCVW60793.2023.00146},
  url       = {https://mlanthology.org/iccvw/2023/honig2023iccvw-biencoder/}
}