stjordanis / modals

MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space

Update (20 Jan 2020): MODALS on text data is avialable

MODALS

MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space

Table of Contents

1. Introduction
2. Getting Started
3. Run Search
4. Run Training
5. Citation

Introduction

MODALS is a framework to apply automated data augmentation to augment data for any modality in a generic way. It exploits automated data augmentation to fine-tune four universal data transformation operations in the latent space to adapt the transform to data of different modalities.

This repository contains code for the work "MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space" (https://openreview.net/pdf?id=XjYgR6gbCEc) implemented using the PyTorch library. It includes searching and training of the SST2 and TREC6 datasets.

Getting Started

Code supports Python 3.

Install requirements

pip install -r requirements.txt

Setting up directory path

In modals/setup.py, specify the dataset path for DATA_DIR and the path to the directory that contains the glove embeddings for EMB_DIR.

Run MODALS search

Script to search for the augmentation policy for SST2 and TREC6 datasets is located in scripts/search.sh. Pass the dataset name as the arguement to call the script.

For example, to search for the augmentation policy for SST2 dataset:

bash scripts/search.sh sst2

The training log and candidate policies of the search will be output to the ./ray_experiments directory.

Run MODALS training

Two searched policy is included in the ./schedule directory. The script to apply the searched policy for training SST2 and TREC6 is located in scripts/train.sh. Pass the dataset name as the arguement to call the script.

bash scripts/train.sh sst2

Citation

If you use MODALS in your research, please cite:

@inproceedings{cheung2021modals,
  title     =  {{\{}MODALS{\}}: Modality-agnostic Automated Data Augmentation in the Latent Space},
  author    =  {Tsz-Him Cheung and Dit-Yan Yeung},
  booktitle =  {International Conference on Learning Representations},
  year      =  {2021},
  url       =  {https://openreview.net/forum?id=XjYgR6gbCEc}
}