@inproceedings{dikhtiar-etal-2026-abateam,
    title = "abateam at {S}em{E}val-2026 Task 1: Plan2joke {--} Humor Policies for Type-Specific Two-Pass Humor Generation",
    author = "Dikhtiar, Andrii  and
      Viter, Antonii  and
      Karaziia, Bohdan  and
      Dementieva, Daryna  and
      Fraser, Alexander",
    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-1.295/",
    pages = "2330--2346",
    ISBN = "979-8-89176-414-9",
    abstract = "Our work was inspired by several recent directions in computational humor and evaluation, including:- Baranov, Kniazhevsky, and Braslavski, ``You Told Me That Joke Twice: A Systematic Investigation of Transferability and Robustness of Humor Detection Models'' (2023).- Tikhonov and Shtykovskiy, ``Humor Mechanics: Advancing Humor Generation with Multistep Reasoning'' (2024).- Zhong, Huang, Gao, Wen, Lin, Zitnik, and Zhou, ``Let{'}s Think Outside the Box: Exploring Leap-of-Thought in Large Language Models with Creative Humor Generation'' (2023).At a high level, we designed a policy-driven humor generation approach covering multiple humor types. We used optimal humor recognition systems and a context enrichment strategy, as well as SFT training based on a dataset composed from previous research samples and adjusted for alignment with our humor policies. This allowed us to perform an ablation study of the approach and to calibrate our system."
}