@inproceedings{kim-etal-2026-d3,
    title = "D3: Dynamic Docid Decoding for Multi-Intent Generative Retrieval",
    author = "Kim, Jaeyoung  and
      Lee, Dohyeon  and
      Hong, Soona  and
      Hwang, Seung-won",
    editor = {Matusevych, Yevgen  and
      Eryi{\u{g}}it, G{\"u}l{\c{s}}en  and
      Aletras, Nikolaos},
    booktitle = "Proceedings of the 19th Conference of the {E}uropean Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 5: Industry Track)",
    month = mar,
    year = "2026",
    address = "Rabat, Morocco",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-eacl/2026.eacl-industry.58/",
    pages = "789--800",
    ISBN = "979-8-89176-384-5",
    abstract = "Generative Retrieval (GR) maps queries to documents by generating discrete identifiers (DocIDs).However, offline DocID assignment and constrained decoding often prevent GR from capturing query-specific intent, especially when documents express multiple or unseen intents (i.e., intent misalignment).We introduce Dynamic Docid Decoding (D3), an inference-time mechanism that adaptively refines DocIDs through delayed, query-informed identifier expansion.D3 uses (a) verification to detect intent misalignment and (b) dynamic decoding to extend DocIDs with query-aligned tokens, even those absent from the pre-indexed vocabulary, enabling plug-and-play DocID expansion beyond the static vocabulary while adding minimal overhead.Experiments on NQ320k and MS-MARCO show that D3 consistently improves retrieval accuracy, especially on unseen and multi-intent documents, across various GR models, including a +2.4{\%}p nDCG@10 gain on the state-of-the-art model."
}