Math word Problem generation using transformers and reinforcement learning

  • Unique Paper ID: 176812
  • Volume: 11
  • Issue: 11
  • PageNo: 7593-7598
  • Abstract:
  • Manually crafting math word problems is a labour-intensive process that teachers do, and one can sense a growing need for automated systems. However, many of the present models will generate problems that are grammatically correct but semantically incoherent, not solvable, or not aligned with the educational objectives. Addressing these issues is the motivation behind our work that enhances an MWP generation model using transformer architecture and reinforcement learning. Having integrated the topic-expression transformer mechanism, our approach will be to align the problem context with appropriate mathematical operations: MWPs are generated that are linguistically sound and mathematically proper. Towards the future, we would focus on the increase of diversity and complexity of the generated problems and evaluation of model adaptability across different datasets. Finally, we shall end up with an application that is user-friendly to enable real-time generation and interaction with MWPs with improved relevance, solvability and effectiveness in the educational setting.

Copyright & License

Copyright © 2025 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

BibTeX

@article{176812,
        author = {Naresh M and D. Naga Jyothi and P. Guru Prasad},
        title = {Math word Problem generation using transformers and reinforcement learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {11},
        pages = {7593-7598},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=176812},
        abstract = {Manually crafting math word problems is a labour-intensive process that teachers do, and one can sense a growing need for automated systems. However, many of the present models will generate problems that are grammatically correct but semantically incoherent, not solvable, or not aligned with the educational objectives. Addressing these issues is the motivation behind our work that enhances an MWP generation model using transformer architecture and reinforcement learning. Having integrated the topic-expression transformer mechanism, our approach will be to align the problem context with appropriate mathematical operations: MWPs are generated that are linguistically sound and mathematically proper. Towards the future, we would focus on the increase of diversity and complexity of the generated problems and evaluation of model adaptability across different datasets. Finally, we shall end up with an application that is user-friendly to enable real-time generation and interaction with MWPs with improved relevance, solvability and effectiveness in the educational setting.},
        keywords = {Math Word Problems (MWPs) · Automated Problem Generation · Transformer Architecture · Reinforcement Learning · Natural Language Processing (NLP).},
        month = {April},
        }

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 11
  • PageNo: 7593-7598

Math word Problem generation using transformers and reinforcement learning

Related Articles