@inproceedings{tam-etal-2026-context,
    title = "The Context Trap: Why End-to-End Audio Language Models Fail Multi-turn Dialogues",
    author = "Tam, Zhi Rui  and
      Chang, Wen Yu  and
      Chen, Yun-Nung",
    editor = "Riccardi, Giuseppe  and
      Mousavi, Seyed Mahed  and
      Torres, Maria Ines  and
      Yoshino, Koichiro  and
      Callejas, Zoraida  and
      Chowdhury, Shammur Absar  and
      Chen, Yun-Nung  and
      Bechet, Frederic  and
      Gustafson, Joakim  and
      Damnati, G{\'e}raldine  and
      Papangelis, Alex  and
      D{'}Haro, Luis Fernando  and
      Mendon{\c{c}}a, John  and
      Bernardi, Raffaella  and
      Hakkani-Tur, Dilek  and
      Di Fabbrizio, Giuseppe {''}Pino{''}  and
      Kawahara, Tatsuya  and
      Alam, Firoj  and
      Tur, Gokhan  and
      Johnston, Michael",
    booktitle = "Proceedings of the 16th International Workshop on Spoken Dialogue System Technology",
    month = feb,
    year = "2026",
    address = "Trento, Italy",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/dashboard-stats/2026.iwsds-1.7/",
    pages = "76--82",
    abstract = "This study systematically compares end-to-end ({E}2{E}) audio language models ({A}udio{LM}s) against modular ({ASR}, {LLM}, {TTS}) systems for multi-phase task-oriented dialogues. We evaluate open-source models on key metrics: conversational naturalness and dialogue consistency. Our findings show that {E}2{E} configurations consistently underperform their modular counterparts, exhibiting severe degradation in dialogue quality across turns. Investigating this failure, our analysis reveals that the core issue lies in the {E}2{E} models' dialogue modeling capabilities, specifically in context maintenance and topic tracking. This work highlights a critical gap between the purported low-latency benefit of {A}udio{LM}s and their practical ability to maintain coherence in complex, multi-turn dialogues, suggesting a need for focused architectural improvements."
}