DARR: A Dual-branch Arithmetic Regression Reasoning Framework for Solving Machine Number Reasoning

This is the official implementation of our AAAI 2025 Oral paper:
DARR: A Dual-Branch Arithmetic Regression Reasoning Framework for Solving Machine Number Reasoning
Chengtai Li*, Yee Yang Tan*, Yuting He, Jianfeng Ren, Ruibin Bai, Yitian Zhao, Heng Yu, Xudong Jiang
Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2025.
[Video] [Poster]

Machine Number Reasoning (MNR) Dataset

Main Results

Requirements

For machine number reasoning (MNR) dataset:

• Python 2.7
• OpenCV
• See mnr_dataset/requirements.txt for a detailed list of packages required.

Experiments

Model training and evaluation code will be released soon.

Citation

If you find this repo useful in your research, please consider citing our paper as follows:

@inproceedings{li2025darr,
  title={DARR: A dual-branch arithmetic regression reasoning framework for solving machine number reasoning},
  author={Li, Chengtai and Tan, Yee Yang and He, Yuting and Ren, Jianfeng and Bai, Ruibin and Zhao, Yitian and Yu, Heng and Jiang, Xudong},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={39},
  number={2},
  pages={1373--1382},
  year={2025}
}

Acknowledgement

We sincerely appreciate the following github repos a lot for their valuable code base: https://github.com/zwh1999anne/Machine-Number-Sense-Dataset