A Deep Reinforcement Learning Approach to Balance Viewport Prediction and Video Transmission in 360° Video Streaming
Abstract
Parameterized Quantum Circuit (PQC) is a family of structured quantum circuits that consists of quantum gates whose parameters are optimized with classical computers. With the quest for a potential speedup, there is a need to run larger quantum circuits, which in turn results in the arduous task of parameter optimization. In this paper, we propose a generic method, called Rotolasso, that utilizes sparsity-inducing coordinate descent (CD) to optimize parameters of a PQC for balancing its accuracy and the number of parameterized gates. The use of CD allows significant reduction in the number of quantum circuit runs, and the sparsity in the model leads to simpler and faster PQCs, both of which are important ingredients to overcome limitations of near-term quantum devices. We provide theoretical analyses and demonstrate experiments showing the effectiveness of Rotolasso to solve instances of combinatorial optimization problems.
Cite
Text
Zhang and Guo. "A Deep Reinforcement Learning Approach to Balance Viewport Prediction and Video Transmission in 360° Video Streaming." International Joint Conference on Artificial Intelligence, 2024. doi:10.24963/ijcai.2024/680Markdown
[Zhang and Guo. "A Deep Reinforcement Learning Approach to Balance Viewport Prediction and Video Transmission in 360° Video Streaming." International Joint Conference on Artificial Intelligence, 2024.](https://mlanthology.org/ijcai/2024/zhang2024ijcai-deep/) doi:10.24963/ijcai.2024/680BibTeX
@inproceedings{zhang2024ijcai-deep,
title = {{A Deep Reinforcement Learning Approach to Balance Viewport Prediction and Video Transmission in 360° Video Streaming}},
author = {Zhang, Guanghui and Guo, Jing},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2024},
pages = {6152-6160},
doi = {10.24963/ijcai.2024/680},
url = {https://mlanthology.org/ijcai/2024/zhang2024ijcai-deep/}
}