@inproceedings{tang-etal-2025-modeling,
    title = "Modeling Uncertainty in Composed Image Retrieval via Probabilistic Embeddings",
    author = "Tang, Haomiao  and
      Wang, Jinpeng  and
      Peng, Yuang  and
      Meng, GuangHao  and
      Luo, Ruisheng  and
      Chen, Bin  and
      Chen, Long  and
      Wang, Yaowei  and
      Xia, Shu-Tao",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.61/",
    pages = "1210--1222",
    ISBN = "979-8-89176-251-0",
    abstract = "Composed Image Retrieval (CIR) enables users to search for images using multimodal queries that combine text and reference images. While metric learning methods have shown promise, they rely on deterministic point embeddings that fail to capture the inherent uncertainty in the input data, in which user intentions may be imprecisely specified or open to multiple interpretations. We address this challenge by reformulating CIR through our proposed Composed Probabilistic Embedding (CoPE) framework, which represents both queries and targets as Gaussian distributions in latent space rather than fixed points. Through careful design of probabilistic distance metrics and hierarchical learning objectives, CoPE explicitly captures uncertainty at both instance and feature levels, enabling more flexible, nuanced, and robust matching that can handle polysemy and ambiguity in search intentions. Extensive experiments across multiple benchmarks demonstrate that CoPE effectively quantifies both quality and semantic uncertainties within Composed Image Retrieval, achieving state-of-the-art performance on recall rate. Code: https://github.com/tanghme0w/ACL25-CoPE."
}