Math Word Problem Reasoning Survey

This repository is a reading list on Mathematical Reasoning, including datasets, methods, and models.

• **** [paper] [code] [dataset]

👀 Table of Contents

• 📊 Benchmarks
• 🚀 Models
• 🔍 Methods

Benchmarks

Vietnamese

• Translate Dataset [dataset]

English & Chinese

• Training Verifiers to Solve Math Word Problems, GSM8K [paper] [code] [dataset]
• Measuring Mathematical Problem Solving With the MATH Dataset, NeurIPS 2021 [paper] [code] [dataset]
• **** [paper] [code] [dataset]

Low Resource

• Analyzing Korean Math Word Problem Data Classification Difficulty Level Using the KoEPT Model, Koreascience [paper]
• Synthetic Data Generator for Solving Korean Arithmetic Word Problem, MDPI [paper] [code]
• BMWP: The First Bengali Math Word Problems Dataset for Operation Prediction and Solving, Under Review [paper]
• Dataset for Evaluation of Mathematical Reasoning Abilities in Russian, AINL [paper] [code]
• ArMATH: a Dataset for Solving Arabic Math Word Problems, LREC [paper] [code]
• WARM: A Weakly (+Semi) Supervised Math Word Problem Solver, Hindi, ICCL 2022 [paper] [code]
• HAWP: a Dataset for Hindi Arithmetic Word Problem Solving, LREC 2022 [paper] [code]

Multilingual

• LANGUAGE MODELS ARE MULTILINGUAL CHAIN-OF-THOUGHT REASONERS [paper] [code]
  • The 10 languages are: Spanish, French, German, Russian, Chinese, Japanese, Thai, Swahili, Bengali, Telugu.

Models

Combining Tool

• NuminaMath 7B TIR [paper] [code]
• MuMath-Code: Combining Tool-Use Large Language Models with Multi-perspective Data Augmentation for Mathematical Reasoning [paper]
• **** [paper] [code] [dataset]

COT

• MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Models, ICLR 2024 [paper] [code] [page]
• Orca-Math: Unlocking the Potential of SLMs in Grade School Math, Microsoft [paper] [summary]
• 🦣 MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuning, ICLR 2024 [paper] [code]
• 🦣 MAmmoTH2: Scaling Instructions from the Web [paper] [code] [page]
• MathChat: Benchmarking Mathematical Reasoning and Instruction Following in Multi-Turn Interactions [paper] [code]
• LLEMMA: AN OPEN LANGUAGE MODEL FOR MATHEMATICS [paper] [code]
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models [paper] [code]
• **** [paper] [code] [dataset]

POT

• **** [paper] [code] [dataset]
• **** [paper] [code] [dataset]

COT + POT

• **** [paper] [code] [dataset]
• WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct [paper] [code] [WizardLM2]

Methods

Curations, collections and surveys

• GitHub: AI for Mathematics (AI4Math) Papers
• GitHub: Awesome LLM4Math
• GitHub: Resources of deep learning for mathematical reasoning (DL4MATH).