@inproceedings{zhang-etal-2026-discal,
    title = "{D}is{C}al: Distribution-Aware Calibration for Mathematical Reasoning Under Character-Level Noisy Inputs",
    author = "Zhang, Bo  and
      Zhang, Jiawei  and
      Gao, Cong  and
      Han, Bingxu  and
      Hu, Minghao  and
      Zhang, Jun  and
      Cao, Yunbo  and
      Luo, Zhunchen  and
      Yao, Wen  and
      Geng, Guotong  and
      Wang, Zhong",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.660/",
    pages = "14484--14507",
    ISBN = "979-8-89176-390-6",
    abstract = "Although large reasoning models (LRMs) exhibit exceptional mathematical reasoning capabilities on clean inputs, their reasoning accuracy drops substantially in the presence of character-level noise such as typographical errors. Critically, their confidence estimates fail to reflect the corresponding decline in reasoning accuracy. While confidence calibration offers a principled solution, existing methods predominantly target clean inputs, leaving noisy scenarios largely unexplored. To address this gap, we propose DisCal (Distribution-aware Calibration), a confidence calibration framework for character-level noisy inputs. DisCal extracts uncertainty signals from both the empirical answer distribution and the model{'}s predictive distribution, and integrates them via a learned calibrator to produce well-calibrated confidence. Experiments across multiple mathematical reasoning benchmarks demonstrate that DisCal consistently outperforms existing calibration methods under noisy inputs, reducing Expected Calibration Error (ECE) by up to 39.21{\%} and improving Area Under the Receiver Operating Characteristic Curve (AUROC) by up to 31.44{\%}."
}