@inproceedings{lu-etal-2024-mathgenie,
title = "{M}ath{G}enie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of {LLM}s",
author = "Lu, Zimu and
Zhou, Aojun and
Ren, Houxing and
Wang, Ke and
Shi, Weikang and
Pan, Junting and
Zhan, Mingjie and
Li, Hongsheng",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = aug,
year = "2024",
address = "Bangkok, Thailand",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.luhme-long.151/",
doi = "10.18653/v1/2024.acl-long.151",
pages = "2732--2747",
abstract = "Large language models (LLMs) have exhibited great potential in mathematical reasoning. However, there remains a performance gap in this area between existing open-source models and closed-source models such as GPT-4. In this paper, we introduce MathGenie, a novel method for generating diverse and reliable math problems by leveraging the ground-truth solutions of the seed data. We augment these ground-truth solutions and use a specially finetuned model to translate these augmented solutions back into new questions. Subsequently, we generate code-integrated solutions for these questions. To ensure the correctness of the code-integrated solutions, we employ rationale-based verification for filtering. Then, we finetune various pretrained models, ranging from 7B to 70B, on the newly curated data, resulting in a family of models known as MathGenie. These models consistently outperform previous open-source models across five representative mathematical reasoning datasets, achieving state-of-the-art performance. In particular, MathGenie-InternLM2 achieves an accuracy of 87.7{\%} on GSM8K and 55.7{\%} on MATH, securing the best overall score."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lu-etal-2024-mathgenie">
<titleInfo>
<title>MathGenie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zimu</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aojun</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Houxing</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ke</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weikang</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junting</namePart>
<namePart type="family">Pan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mingjie</namePart>
<namePart type="family">Zhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongsheng</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lun-Wei</namePart>
<namePart type="family">Ku</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Martins</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="family">Srikumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Bangkok, Thailand</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Large language models (LLMs) have exhibited great potential in mathematical reasoning. However, there remains a performance gap in this area between existing open-source models and closed-source models such as GPT-4. In this paper, we introduce MathGenie, a novel method for generating diverse and reliable math problems by leveraging the ground-truth solutions of the seed data. We augment these ground-truth solutions and use a specially finetuned model to translate these augmented solutions back into new questions. Subsequently, we generate code-integrated solutions for these questions. To ensure the correctness of the code-integrated solutions, we employ rationale-based verification for filtering. Then, we finetune various pretrained models, ranging from 7B to 70B, on the newly curated data, resulting in a family of models known as MathGenie. These models consistently outperform previous open-source models across five representative mathematical reasoning datasets, achieving state-of-the-art performance. In particular, MathGenie-InternLM2 achieves an accuracy of 87.7% on GSM8K and 55.7% on MATH, securing the best overall score.</abstract>
<identifier type="citekey">lu-etal-2024-mathgenie</identifier>
<identifier type="doi">10.18653/v1/2024.acl-long.151</identifier>
<location>
<url>https://aclanthology.org/2024.luhme-long.151/</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>2732</start>
<end>2747</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MathGenie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of LLMs
%A Lu, Zimu
%A Zhou, Aojun
%A Ren, Houxing
%A Wang, Ke
%A Shi, Weikang
%A Pan, Junting
%A Zhan, Mingjie
%A Li, Hongsheng
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand
%F lu-etal-2024-mathgenie
%X Large language models (LLMs) have exhibited great potential in mathematical reasoning. However, there remains a performance gap in this area between existing open-source models and closed-source models such as GPT-4. In this paper, we introduce MathGenie, a novel method for generating diverse and reliable math problems by leveraging the ground-truth solutions of the seed data. We augment these ground-truth solutions and use a specially finetuned model to translate these augmented solutions back into new questions. Subsequently, we generate code-integrated solutions for these questions. To ensure the correctness of the code-integrated solutions, we employ rationale-based verification for filtering. Then, we finetune various pretrained models, ranging from 7B to 70B, on the newly curated data, resulting in a family of models known as MathGenie. These models consistently outperform previous open-source models across five representative mathematical reasoning datasets, achieving state-of-the-art performance. In particular, MathGenie-InternLM2 achieves an accuracy of 87.7% on GSM8K and 55.7% on MATH, securing the best overall score.
%R 10.18653/v1/2024.acl-long.151
%U https://aclanthology.org/2024.luhme-long.151/
%U https://doi.org/10.18653/v1/2024.acl-long.151
%P 2732-2747
Markdown (Informal)
[MathGenie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of LLMs](https://aclanthology.org/2024.luhme-long.151/) (Lu et al., ACL 2024)
ACL
- Zimu Lu, Aojun Zhou, Houxing Ren, Ke Wang, Weikang Shi, Junting Pan, Mingjie Zhan, and Hongsheng Li. 2024. MathGenie: Generating Synthetic Data with Question Back-translation for Enhancing Mathematical Reasoning of LLMs. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2732–2747, Bangkok, Thailand. Association for Computational Linguistics.