Summary of Energy-efficient Power Control For Multiple-task Split Inference in Uavs: a Tiny Learning-based Approach, by Chenxi Zhao et al.
Energy-Efficient Power Control for Multiple-Task Split Inference in UAVs: A Tiny Learning-Based Approach
by Chenxi Zhao, Min Sheng, Junyu Liu, Tianshu Chu, Jiandong Li
First submitted to arxiv on: 31 Dec 2023
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Multiagent Systems (cs.MA); Robotics (cs.RO)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a two-timescale approach to minimize energy consumption in split inference for unmanned aerial vehicles (UAVs). The method segregates discrete and continuous variables into two timescales, reducing the size of the action space and computational complexity. Tiny reinforcement learning (TRL) is used to select transmission modes for sequential tasks, while optimization programming (OP) optimizes transmit power. The algorithm replaces optimizing transmit power with transmission time, which reduces energy consumption and enables a fast solution. Simulation results show that this approach achieves higher task completion probability with lower energy consumption. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary For UAVs, aerial artificial intelligence is limited by energy and computing resources. To solve this problem, researchers use split inference. However, achieving energy-efficient split inference in UAVs is complex because of many factors to consider. This paper proposes a new way to do it. The method uses two timescales: one for discrete variables and one for continuous variables. This makes the calculations easier and faster. The approach also combines tiny reinforcement learning (TRL) with optimization programming (OP) to optimize transmission power. In this case, the OP problem is simplified by replacing transmit power optimization with transmission time optimization. This reduces energy consumption and makes the algorithm faster. Simulation results show that this approach works well. |
Keywords
* Artificial intelligence * Inference * Optimization * Probability * Reinforcement learning
