@inproceedings{ke-etal-2025-catch,
    title = "{CATCH}: A Controllable Theme Detection Framework with Contextualized Clustering and Hierarchical Generation",
    author = "Ke, Rui  and
      Xu, Jiahui  and
      Wang, Kuang  and
      Yang, Shenghao  and
      Jiang, Feng  and
      Li, Haizhou",
    editor = "Hedayatnia, Behnam  and
      Chen, Vivian  and
      Chen, Zhang  and
      Gupta, Raghav  and
      Galley, Michel",
    booktitle = "Proceedings of the Twelfth Dialog System Technology Challenge",
    month = aug,
    year = "2025",
    address = "Avignon, France",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/corrections-2025-10/2025.dstc-1.2/",
    pages = "17--26",
    ISBN = "979-8-89176-330-2",
    abstract = "Theme detection is a fundamental task in user-centric dialogue systems, aiming to identify the latent topic of each utterance without relying on predefined schemas. Unlike intent induction, which operates within fixed label spaces, theme detection requires cross-dialogue consistency and alignment with personalized user preferences, posing significant challenges. Existing methods often struggle with sparse, short utterances and fail to capture user-level thematic preferences across dialogues. To address these challenges, we propose CATCH (Controllable Theme Detection with Contextualized Clustering and Hierarchical Generation), a unified framework that integrates three core components: (1) context-aware topic representation, which enriches utterance-level semantics using surrounding topic segments; (2) preference-guided topic clustering, which jointly models semantic proximity and personalized feedback to align themes across conversations; and (3) a hierarchical theme generation mechanism designed to suppress noise and produce robust, coherent topic labels. Experiments on a multi-domain customer dialogue benchmark demonstrate that CATCH achieves state-of-the-art performance in both theme classification and topic distribution quality. Notably, it ranked second in the official blind evaluation of the DSTC-12 Controllable Theme Detection Track, showcasing its effectiveness and generalizability in real-world dialogue systems."
}