A Tree-Structured Multi-Task Model Recommender

Lijun Zhang, Xiao Liu, Hui Guan
Proceedings of the First International Conference on Automated Machine Learning, PMLR 188:10/1-12, 2022.

Abstract

Tree-structured multi-task architectures have been employed to jointly tackle multiple vision tasks in the context of multi-task learning (MTL). The major challenge is to determine where to branch out for each task given a backbone model to optimize for both task accuracy and computation efficiency. To address the challenge, this paper proposes a recommender that, given a set of tasks and a convolutional neural network-based backbone model, automatically suggests tree-structured multi-task architectures that could achieve a high task performance while meeting a user-specified computation budget without performing model training. Extensive evaluations on popular MTL benchmarks show that the recommended architectures could achieve competitive task accuracy and computation efficiency compared with state-of-the-art MTL methods. Our tree-structured multi-task model recommender is open-sourced and available at \url{https://github.com/zhanglijun95/TreeMTL}.

Cite this Paper

BibTeX
@InProceedings{pmlr-v188-zhang22a, title = {A Tree-Structured Multi-Task Model Recommender}, author = {Zhang, Lijun and Liu, Xiao and Guan, Hui}, booktitle = {Proceedings of the First International Conference on Automated Machine Learning}, pages = {10/1--12}, year = {2022}, editor = {Guyon, Isabelle and Lindauer, Marius and van der Schaar, Mihaela and Hutter, Frank and Garnett, Roman}, volume = {188}, series = {Proceedings of Machine Learning Research}, month = {25--27 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v188/zhang22a/zhang22a.pdf}, url = {https://proceedings.mlr.press/v188/zhang22a.html}, abstract = {Tree-structured multi-task architectures have been employed to jointly tackle multiple vision tasks in the context of multi-task learning (MTL). The major challenge is to determine where to branch out for each task given a backbone model to optimize for both task accuracy and computation efficiency. To address the challenge, this paper proposes a recommender that, given a set of tasks and a convolutional neural network-based backbone model, automatically suggests tree-structured multi-task architectures that could achieve a high task performance while meeting a user-specified computation budget without performing model training. Extensive evaluations on popular MTL benchmarks show that the recommended architectures could achieve competitive task accuracy and computation efficiency compared with state-of-the-art MTL methods. Our tree-structured multi-task model recommender is open-sourced and available at \url{https://github.com/zhanglijun95/TreeMTL}.} }
Endnote
%0 Conference Paper %T A Tree-Structured Multi-Task Model Recommender %A Lijun Zhang %A Xiao Liu %A Hui Guan %B Proceedings of the First International Conference on Automated Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Isabelle Guyon %E Marius Lindauer %E Mihaela van der Schaar %E Frank Hutter %E Roman Garnett %F pmlr-v188-zhang22a %I PMLR %P 10/1--12 %U https://proceedings.mlr.press/v188/zhang22a.html %V 188 %X Tree-structured multi-task architectures have been employed to jointly tackle multiple vision tasks in the context of multi-task learning (MTL). The major challenge is to determine where to branch out for each task given a backbone model to optimize for both task accuracy and computation efficiency. To address the challenge, this paper proposes a recommender that, given a set of tasks and a convolutional neural network-based backbone model, automatically suggests tree-structured multi-task architectures that could achieve a high task performance while meeting a user-specified computation budget without performing model training. Extensive evaluations on popular MTL benchmarks show that the recommended architectures could achieve competitive task accuracy and computation efficiency compared with state-of-the-art MTL methods. Our tree-structured multi-task model recommender is open-sourced and available at \url{https://github.com/zhanglijun95/TreeMTL}.
APA
Zhang, L., Liu, X. & Guan, H.. (2022). A Tree-Structured Multi-Task Model Recommender. Proceedings of the First International Conference on Automated Machine Learning, in Proceedings of Machine Learning Research 188:10/1-12 Available from https://proceedings.mlr.press/v188/zhang22a.html.

Related Material