%0 Conference Paper %1 Fang2017DynamicML %A Fang, Yuchun %A Ma, Zhengyan %A Zhang, Zhaoxiang %A Zhang, Xu-Yao %A Bai, Xiang %B IJCAI %D 2017 %K deep_learning multitask vision %T Dynamic Multi-Task Learning with Convolutional Neural Network %X Multi-task learning and deep convolutional neural network (CNN) have been successfully used in var- ious fields. This paper considers the integration of CNN and multi-task learning in a novel way to fur- ther improve the performance of multiple related tasks. Existing multi-task CNN models usually em- pirically combine different tasks into a group which is then trained jointly with a strong assumption of model commonality. Furthermore, traditional ap- proaches usually only consider small number of tasks with rigid structure, which is not suitable for large-scale applications. In light of this, we propose a dynamic multi-task CNN model to handle these problems. The proposed model directly learns the task relations from data instead of subjective task grouping. Due to its flexible structure, it supports task-wise incremental training, which is useful for efficient training of massive tasks. Specifically, we add a new task transfer connection (TTC) between the layers of each task. The learned TTC is able to reflect the correlation among different tasks guid- ing the model dynamically adjusting the multiplex- ing of the information among different tasks. With the help of TTC, multiple related tasks can further boost the whole performance for each other. Ex- periments demonstrate that the proposed dynamic multi-task CNN model outperforms traditional ap- proaches.

@inproceedings{Fang2017DynamicML, abstract = {Multi-task learning and deep convolutional neural network (CNN) have been successfully used in var- ious fields. This paper considers the integration of CNN and multi-task learning in a novel way to fur- ther improve the performance of multiple related tasks. Existing multi-task CNN models usually em- pirically combine different tasks into a group which is then trained jointly with a strong assumption of model commonality. Furthermore, traditional ap- proaches usually only consider small number of tasks with rigid structure, which is not suitable for large-scale applications. In light of this, we propose a dynamic multi-task CNN model to handle these problems. The proposed model directly learns the task relations from data instead of subjective task grouping. Due to its flexible structure, it supports task-wise incremental training, which is useful for efficient training of massive tasks. Specifically, we add a new task transfer connection (TTC) between the layers of each task. The learned TTC is able to reflect the correlation among different tasks guid- ing the model dynamically adjusting the multiplex- ing of the information among different tasks. With the help of TTC, multiple related tasks can further boost the whole performance for each other. Ex- periments demonstrate that the proposed dynamic multi-task CNN model outperforms traditional ap- proaches.}, added-at = {2018-08-27T22:35:33.000+0200}, author = {Fang, Yuchun and Ma, Zhengyan and Zhang, Zhaoxiang and Zhang, Xu-Yao and Bai, Xiang}, biburl = {https://www.bibsonomy.org/bibtex/2f47c193d1a9611029a7ee91398cdc35d/dallmann}, booktitle = {IJCAI}, description = {Dynamic Multi-Task Learning with Convolutional Neural Network - Semantic Scholar}, interhash = {c4f63a87491484291ff7350364654e1b}, intrahash = {f47c193d1a9611029a7ee91398cdc35d}, keywords = {deep_learning multitask vision}, timestamp = {2018-08-27T22:35:33.000+0200}, title = {Dynamic Multi-Task Learning with Convolutional Neural Network}, year = 2017 }