EUVIMEDEUVIMED entdecken
Crossref

MathSE: Improving Multimodal Mathematical Reasoning via Self-Evolving Iterative Reflection and Reward-Guided Fine-Tuning

Jinhao Chen, Zhen Yang, Jianxin Shi, Tianyu Wo, Jie Tang

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (24) · 20136-20144

DOI: 10.1609/aaai.v40i24.39100 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Jinhao Chen und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in vision-language answering tasks. Despite their strengths, these models often encounter challenges in achieving complex reasoning tasks such as mathematical problem-solving. Previous works have focused on fine-tuning on specialized mathematical datasets. However, these datasets are typically distilled directly from teacher models, which capture only static reasoning patterns and leaving substantial gaps compared to student models. This reliance on fixed teacher-derived datasets not only restricts the model's ability to adapt to novel or more intricate questions that extend beyond the confines of the training data, but also lacks the iterative depth needed for robust generalization. To overcome these limitations, we propose MathSE, a Mathematical Self-Evolving framework for MLLMs. In contrast to traditional one-shot fine-tuning paradigms, MathSE iteratively refines the model through cycles of inference, reflection, and reward-based feedback. Specifically, we leverage iterative fine-tuning by incorporating correct reasoning paths derived from previous-stage inference and integrating reflections from a specialized Outcome Reward Model (ORM). To verify the effectiveness of MathSE, we evaluate it on a suite of challenging benchmarks, demonstrating significant performance gains over backbone models. Notably, our experimental results on MathVL-test surpass the leading open-source multimodal mathematical reasoning model QVQ.

Zitieren

APA

Chen, J., Yang, Z., Shi, J., Wo, T., & Tang, J. (2026). MathSE: Improving Multimodal Mathematical Reasoning via Self-Evolving Iterative Reflection and Reward-Guided Fine-Tuning. Proceedings of the AAAI Conference on Artificial Intelligence, 40(24), 20136-20144. https://doi.org/10.1609/aaai.v40i24.39100

Vancouver

Chen J, Yang Z, Shi J, Wo T, Tang J. MathSE: Improving Multimodal Mathematical Reasoning via Self-Evolving Iterative Reflection and Reward-Guided Fine-Tuning. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(24):20136-20144. doi:10.1609/aaai.v40i24.39100

BibTeX
@article{Chen2026,
  author = {Chen, Jinhao and Yang, Zhen and Shi, Jianxin and Wo, Tianyu and Tang, Jie},
  title = {MathSE: Improving Multimodal Mathematical Reasoning via Self-Evolving Iterative Reflection and Reward-Guided Fine-Tuning},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {24},
  pages = {20136--20144},
  doi = {10.1609/aaai.v40i24.39100}
}
Kostenlos starten →