@inproceedings{chen-xuan-qi-etal-2026-trac,
    title = "{TRAC}: Token-level Reward Assignment for Coherent Abstractive Summarization",
    author = "陈宣齐  and
      容梓莹  and
      Liao, Xinfeng  and
      Wang, Lianxi  and
      Gao, Ying  and
      Jiang, Shengyi",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.findings-acl.1873/",
    pages = "37563--37578",
    ISBN = "979-8-89176-395-1",
    abstract = "Large Language Models (LLMs) have achieved remarkable success in text summarization, particularly through the integration of reinforcement learning. However, maintaining logical coherence and contextual consistency remains a pervasive challenge in long-form generation, often hindering the production of high-quality, unified summaries. To address these persistent issues, we propose TRAC, a framework that introduces a token-level reward function by integrating relative sentence gain, inter-sentence attention, and a Gaussian length penalty. By training a Process Reward Model (PRM) to provide fine-grained, step-wise supervision, TRAC ensures superior structural integrity and fluency during the generation process. Experimental results demonstrate that TRAC outperforms the sequence-level baseline by 11.05{\%} in Fluency and 10.61{\%} in Relevance. Furthermore, it achieves significant gains over competitive baselines such as FIGA and TLCR, underscoring its effectiveness and generalizability in high-quality NLP summarization."
}