DenseMTL: Cross-task Attention Mechanism for Dense Multi-task Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lopes, Ivan, Vu, Tuan-Hung, de Charette, Raoul
Natura: Preprint
Pubblicazione: 2022
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910639203549184
author Lopes, Ivan
Vu, Tuan-Hung
de Charette, Raoul
author_facet Lopes, Ivan
Vu, Tuan-Hung
de Charette, Raoul
contents Multi-task learning has recently emerged as a promising solution for a comprehensive understanding of complex scenes. In addition to being memory-efficient, multi-task models, when appropriately designed, can facilitate the exchange of complementary signals across tasks. In this work, we jointly address 2D semantic segmentation and three geometry-related tasks: dense depth estimation, surface normal estimation, and edge estimation, demonstrating their benefits on both indoor and outdoor datasets. We propose a novel multi-task learning architecture that leverages pairwise cross-task exchange through correlation-guided attention and self-attention to enhance the overall representation learning for all tasks. We conduct extensive experiments across three multi-task setups, showing the advantages of our approach compared to competitive baselines in both synthetic and real-world benchmarks. Additionally, we extend our method to the novel multi-task unsupervised domain adaptation setting. Our code is available at https://github.com/cv-rits/DenseMTL
format Preprint
id arxiv_https___arxiv_org_abs_2206_08927
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle DenseMTL: Cross-task Attention Mechanism for Dense Multi-task Learning
Lopes, Ivan
Vu, Tuan-Hung
de Charette, Raoul
Computer Vision and Pattern Recognition
Artificial Intelligence
Robotics
Multi-task learning has recently emerged as a promising solution for a comprehensive understanding of complex scenes. In addition to being memory-efficient, multi-task models, when appropriately designed, can facilitate the exchange of complementary signals across tasks. In this work, we jointly address 2D semantic segmentation and three geometry-related tasks: dense depth estimation, surface normal estimation, and edge estimation, demonstrating their benefits on both indoor and outdoor datasets. We propose a novel multi-task learning architecture that leverages pairwise cross-task exchange through correlation-guided attention and self-attention to enhance the overall representation learning for all tasks. We conduct extensive experiments across three multi-task setups, showing the advantages of our approach compared to competitive baselines in both synthetic and real-world benchmarks. Additionally, we extend our method to the novel multi-task unsupervised domain adaptation setting. Our code is available at https://github.com/cv-rits/DenseMTL
title DenseMTL: Cross-task Attention Mechanism for Dense Multi-task Learning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Robotics
url https://arxiv.org/abs/2206.08927