Curriculum-Based Self-Training Makes Better Few-Shot Learners for Data-to-Text Generation

Introduction

Curriculum-Based Self-Training (CBST) utilizes curriculum learning to construct pseudo-labeled data from easy cases to hard ones, and leverages such data into the self-training process at different iterations. You can read our paper for more details. This project is a PyTorch implementation of our work.

Dependencies

Python 3.7
NumPy
PyTorch 1.4.0
Transformers (Huggingface) 3.0.0

Quick Start

NOTE: In order to compute the METEOR scores, please download the required data and put it under the following folder: pycocoevalcap/meteor/data/.

Datasets

Our experiments involve two datasets, i.e., WebNLG and WikiBio. The raw data are from the GitHub repository of KGPT. You can download the pre-processed datasets used in our paper on Google Drive / Tsinghua Cloud. This data folder also contains a file counterfitted_neighbors.json used for word substitution, which originates from the GitHub repository of certified-word-sub.

Training

You can download the model checkpoint of BART provided by Huggingface Transformers, and train the model on two datasets, respectively.

bash finetune_bart_webnlg.sh
bash finetune_bart_wikibio.sh

In the scripts, --output_dir denotes the directory to save the intermediate and final models. --model_path indicates the pre-trained checkpoint used for initialization. --model_path and --tokenizer_path are set to the directory of the downloaded BART checkpoint. You can refer to the codes for the details of other hyper-parameters.

Inference

We also provide the inference scripts to directly acquire the generation results on the test sets.

bash infer_bart_webnlg.sh
bash infer_bart_wikibio.sh

In the scripts, --output_dir denotes the directory of model checkpoints used for inference. The generated results are also saved in this directory.

Citation

@inproceedings{ke2022cbst,
  title     = {Curriculum-Based Self-Training Makes Better Few-Shot Learners for Data-to-Text Generation},
  author    = {Ke, Pei and Ji, Haozhe and Yang, Zhenyu and Huang, Yi and Feng, Junlan and Zhu, Xiaoyan and Huang, Minlie},
  booktitle = {Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, {IJCAI-22}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  pages     = {4178--4184},
  year      = {2022},
}

Please kindly cite our paper if this paper and the codes are helpful.

Thanks

Many thanks to the GitHub repositories of Transformers and bart-closed-book-qa. Part of our codes are modified based on their codes.

kepei1106/CBST