Semantic Communications With AI Tasks

• DOI: 10.48550/arxiv.2109.14170
• Publication Date: 2021-01-01
• Authors: Yang Yang; Caili Guo; Fangfang Liu; Chuanhong Liu; Lunan Sun; Qizheng Sun; Jiujiu Chen
• Keywords: FOS: Computer and information sciences; 0203 mechanical engineering; Computer Vision and Pattern Recognition (cs.CV); Computer Science - Computer Vision and Pattern Recognition; 02 engineering and technology

A radical paradigm shift of wireless networks from ``connected things'' to intelligence'' undergoes, which coincides with the Shanno and Weaver's envisions: Communications will transform technical level semantic level. This article proposes a communication method artificial intelligence tasks (SC-AIT). First, architecture SC-AIT is elaborated. Then, based on proposed architecture, we implement for image classifications task. prototype also established surface defect detection, conducted. Experimental results show that has much lower bandwidth requirements, can achieve more than $40\%$ classification accuracy gains compared communications at Future trends key challenges are identified.