A Multilingual Dataset (MultiMWP) and Benchmark for Math Word Problem Generation
We present a multi-way parallel corpus of Math Word Problems (MWPs) in nine languages, including six low-resource languages. To date, this is the largest multilingual MWP dataset available. We utilize this dataset and show the viability of using pre-trained multilingual sequence-sequence language models (prMSLMs) for autoregressive MWP generation in both monolingual and multilingual setups, particularly for low-resource languages. We also integrate a math constraint satisfaction module with autoregressive text generation. Our extensive evaluations identify several factors that affect autoregressive text generation on prMSLMs. These include language representation in the model, model size, existence of similar languages in the model, and language script. Overall, our results reveal that autoregressive MWP generation on top of prMSLMs is very promising, even for low-resource languages.
Authors
- Qiyue Zhang (ORCID: https://orcid.org/0000-0003-1956-7093)
- Surangika Ranathunga (ORCID: https://orcid.org/0000-0003-0701-0204)
- V. K. Dwivedi (ORCID: https://orcid.org/0000-0002-0801-4744)
- Dhrubajyoti Kalita
- Rajkumari Monimala Sinha
- Nitisha Aggarwal (ORCID: https://orcid.org/0000-0001-7468-8395)
- Jingwen Ji (ORCID: https://orcid.org/0009-0007-0556-6272)
- Barga Deori
- Aryaveer Singh
- Yuchen Zeng (ORCID: https://orcid.org/0000-0001-7756-2309)
- Omega Gamage (ORCID: https://orcid.org/0009-0004-7074-2699)
- Annie Lee (ORCID: https://orcid.org/0000-0003-3726-6024)
- Marjana Prifti Skënduli (ORCID: https://orcid.org/0000-0002-2707-1621)
- Ananya Ananya
- Ajit Kumar Nayak (ORCID: https://orcid.org/0000-0003-2302-9458)
- Xiao Sun (ORCID: https://orcid.org/0000-0001-9750-7032)
- Mehreen Alam (ORCID: https://orcid.org/0000-0001-7663-2598)
- Yanke Mao
- Vijaykant Pandey
- Haonan Gao
- N. Hussain
- Mu-qiu Tian
- Endi Trico
- Danja Nako
- Sara Hoxha
- Dezi Imami
- Dea Doksani
- Sonila Shqezi
Institutions
- Guru Gobind Singh Indraprastha University (IN)
- University of Delhi (IN)
- Dr. Bhim Rao Ambedkar University (IN)
- National Institute Of Technology Silchar (IN)
- University of Toronto (CA)
- Siksha O Anusandhan University (IN)
- National University of Computer and Emerging Sciences (PK)
- Cotton University (IN)
- Indian Institute of Technology Bhilai (IN)
- Film Independent (US)
- International Center for Biosaline Agriculture (AE)
- Massey University (NZ)
- Indian Institute of Technology BHU (IN)
- University of New York Tirana (AL)
- University of Colombo (LK)
- Banaras Hindu University (IN)
Publication Details
- Journal
- IEEE Transactions on Audio Speech and Language Processing
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1109/taslpro.2025.3552936
- Primary Topic
- Natural Language Processing Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00