AS-ES-learning

Implementation for ACL2024 paper "AS-ES Learning: Towards Efficient CoT Learning in Small Models"

Datasets

We use the following datasets in our paper:

• MWP - Distilling Chain-of-Thought Reasoning from code-davinci-002 to FlanT5: the in-context chain-of-thought part of the data
• PET report summary - cPET-11K, a novel dataset we release in this paper (under /data). A collection of PET/CT report data from patients with pancreatic cancer.

Quick Start

• Environment configurations: conda env create -f environment.yml
• Usage Example:
  • for as-es dataset construction, run bash utils/run.sh
  • for training using as-es dataset, run bash script/train.sh
  • for inference, run bash script/test.sh

Citing

If you find our work helpful, feel free to cite our publication -

AS-ES Learning: Towards Efficient CoT Learning in Small Models

@inproceedings{xi-etal-2024-es,
    title = "{AS}-{ES} Learning: Towards efficient {C}o{T} learning in small models",
    author = "Xi, Nuwa  and
      Chen, Yuhan  and
      Zhao, Sendong  and
      Wang, Haochun  and
      GongZhang, GongZhang  and
      Qin, Bing  and
      Liu, Ting",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand and virtual meeting",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.findings-acl.635",
    pages = "10686--10697"
}

If you have any questions, feel free to contact: Nova X