notesum.ai

Published at November 27

Two-Timescale Digital Twin Assisted Model Interference and Retraining over Wireless Network

eess.SP

cs.IT

math.IT

Released Date: November 27, 2024

Authors: Jiayi Cong¹, Guoliang Cheng¹, Changsheng You², Xinyu Huang³, Wen Wu¹

Aff.: ¹Frontier Research Center, Peng Cheng Laboratory, Shenzhen, China; ²The Department of Electronic and Electrical Engineering, Southern University of Science and Technology; ³The Department of Electrical and Computer Engineering, University of Waterloo

Arxiv: http://arxiv.org/abs/2411.18329v1

Parameter	Value
Training samples	40,000
Test samples	10,000
Incremental samples	5,000
The number of BS, $M$	3
The number of MU, $N$	$U(5,10)$
Unit time, $t_{s}$	5ms
Time slot count, $T$	2
Time frame count, $S$	10
Accuracy constraint, $\bar{C}$	0.85
Computation resource, ${\lambda}^{\text{BS}}$ [16]	50 GFLOPs
MU transmission data size, $Q^{t}_{m,n}$	$U(10,20)$ MB
Transmission power, $\omega$	-10 dBm
Transmission noise power, $\sigma^{2}$	-100 dBm