Neural Architecture Search in the Context of Deep Multi-Task Learning

Guilherme Gadelha, Herman Gomes, Leonardo Batista

2023

Abstract

Multi-Task Learning (MTL) is a neural network design paradigm that aims to improve generalization while simultaneously solving multiple tasks. It has obtained success in many application areas such as Natural Language Processing and Computer Vision. In an MTL neural network, there are shared task branches and task-specific branches. However, automatically deciding on the best locations and sizes of those branches as a result of the domain tasks remains an open question. With the aim of shedding light to the above question, we designed a sequence of experiments involving single-task networks, multi-task networks, and networks created with a neural architecture search (NAS) strategy. In addition, we proposed a competitive neural network architecture for a challenging use case: the ICAO photograph conformance checking for issuing of passports. We obtained the best results using a handcrafted MTL network, whose effectiveness is close to state-of-the-art methods. Furthermore, our experiments and analysis pave the way to develop a technique to automatically create branches and group similar tasks into an MTL network.

Paper Citation

in Harvard Style

Gadelha G., Gomes H. and Batista L. (2023). Neural Architecture Search in the Context of Deep Multi-Task Learning. In Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 5: VISAPP; ISBN 978-989-758-634-7, SciTePress, pages 684-691. DOI: 10.5220/0011696200003417

in Bibtex Style

@conference{visapp23,
author={Guilherme Gadelha and Herman Gomes and Leonardo Batista},
title={Neural Architecture Search in the Context of Deep Multi-Task Learning},
booktitle={Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 5: VISAPP},
year={2023},
pages={684-691},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011696200003417},
isbn={978-989-758-634-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 5: VISAPP
TI - Neural Architecture Search in the Context of Deep Multi-Task Learning
SN - 978-989-758-634-7
AU - Gadelha G.
AU - Gomes H.
AU - Batista L.
PY - 2023
SP - 684
EP - 691
DO - 10.5220/0011696200003417
PB - SciTePress