What is it about?
Multi-task Learning (MTL) has yielded immense success in Recommender System (RS) applications. This paper proposes a novel multi-task learning framework, RMTL, to improve the prediction performance of multi-tasks by generating dynamic total loss weights in an RL manner and better capture session-wise information.
Featured Image
Photo by Possessed Photography on Unsplash
Why is it important?
Most existing MTL works typically use linear scalarizations of the multiple-task loss functions and item-wise modeling, which result in limited performance. To address the two above-mentioned problems, we propose an RL-enhanced multi-task recommendation framework, RMTL, which is capable of incorporating the sequential property of user-item interactions into MTL recommendations and automatically updating the task-wise weights in the overall loss function.
Perspectives
This article is the first to apply reinforcement learning to MTL and has achieved good results. The design of loss weights focuses more on training tasks with lower future accuracy which may help to improve the optimization process. I hope it can be thought-provoking and provide inspiration to everyone.
Ziru Liu
City University of Hong Kong
Read the Original
This page is a summary of: Multi-Task Recommendations with Reinforcement Learning, April 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3543507.3583467.
You can read the full text:
Resources
Contributors
The following have contributed to this page
