@inproceedings{kim-etal-2026-clulab,
    title = "clulab-retrieval at {S}em{E}val-2026 Task 8: A Comparative Analysis of Dense Retrievers and {H}y{DE} for Multi-Turn Conversational Retrieval",
    author = "Kim, Hyungji  and
      Kondapaneni, Siva Rohit  and
      Bethard, Steven",
    editor = "Kochmar, Ekaterina  and
      Ghosh, Debanjan  and
      North, Kai  and
      Komachi, Mamoru",
    booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.semeval-1.351/",
    pages = "2787--2792",
    ISBN = "979-8-89176-414-9",
    abstract = "We present a comparative analysis of dense retrievers and retrieval strategies for multi-turn conversational retrieval in SemEval-2026 Task 8 (MTRAGEval). Our official submission employed a fine-tuned E5-based dense retriever (E5-FT, {\textasciitilde}110M parameters) with Hypothetical Document Embeddings (HyDE), achieving nDCG@5 of .3309, ranking 31 out of 38 systems. On the development set we also compared E5-FT versus BGE embeddings, dense-only versus hybrid retrieval strategies, and HyDE versus keyword extraction approaches. We found: (1) BGE (general-purpose, {\textasciitilde}110M) outperforms our domain-fine-tuned E5-FT ({\textasciitilde}110M) by 30.5{\%} on baseline retrieval, suggesting that model selection may matter more than domain-specific fine-tuning, (2) hybrid retrieval combining BM25 and dense methods provides complementary signals, with HyDE improving BM25 by 26.7{\%} and dense retrieval by 4.0{\%}, and (3) keyword-based query simplification degrades performance by 11-28{\%} across domains, validating HyDE{'}s approach of preserving semantic richness through passage-level text."
}