MEDIC: A Multi-Task Learning Dataset for Disaster Image Classification

TL;DR

MEDIC is the largest multi-task social media image dataset for disaster response, enabling research in image-based disaster management and multi-task learning with promising experimental results.

Contribution

This paper introduces MEDIC, the first large-scale multi-task social media image dataset for disaster response, advancing research in image-based disaster management and multi-task learning.

Findings

Deep learning architectures outperform baselines on all tasks.

The dataset facilitates multi-task learning research in disaster informatics.

Promising results demonstrate the dataset's potential for real-world applications.

Abstract

Recent research in disaster informatics demonstrates a practical and important use case of artificial intelligence to save human lives and suffering during natural disasters based on social media contents (text and images). While notable progress has been made using texts, research on exploiting the images remains relatively under-explored. To advance image-based approaches, we propose MEDIC (Available at: https://crisisnlp.qcri.org/medic/index.html), which is the largest social media image classification dataset for humanitarian response consisting of 71,198 images to address four different tasks in a multi-task learning setup. This is the first dataset of its kind: social media images, disaster response, and multi-task learning research. An important property of this dataset is its high potential to facilitate research on multi-task learning, which recently receives much interest from the machine learning community and has shown remarkable results in terms of memory, inference speed, performance, and generalization capability. Therefore, the proposed dataset is an important resource for advancing image-based disaster management and multi-task machine learning research. We experiment with different deep learning architectures and report promising results, which are above the majority baselines for all tasks. Along with the dataset, we also release all relevant scripts (https://github.com/firojalam/medic).