学术报告：（11月29日）A Deep Learning Estimator for the Signal Detection of SIMO system with 1-bit Quantized Output

报告题目：A Deep Learning Estimator for the Signal Detection of SIMO system with 1-bit Quantized Output

报告时间：2018年11月29日上午10:00-11:30

报告地点：最热门的网赌网址大全十大网赌网址信誉排行榜微电子大楼-208

特邀专家：沈聪教授，中国科学技术大学

主持人：陈翔 副教授

专家简介：

Cong Shen received his B.S. and M.S. degrees, in 2002 and 2004 respectively, from the Department of Electronic Engineering, Tsinghua University, China. He obtained the Ph.D. degree from the Electrical Engineering Department, UCLA, in 2009. From 2009 to 2014, He worked for Qualcomm Research in San Diego, CA, In 2015, he returned to academia and joined University of Science and Technology of China (USTC) as the 100 Talents Program Professor in the School of Information Science and Technology. His general research interests are in the area of machine learning, information theory, communication theory, and wireless networks. In particular, his current research focuses on multi-armed bandit and its applications, deep learning, and statistical information processing.

He was the recipient of the 2015 100 Talents Program of the Chinese Academy of Science. In 2017 he received the “Excellent Paper Award” in the 9th International Conference on Ubiquitous and Future Networks (ICUFN 2017). Currently, he serves as an editor for the IEEE Transactions on Wireless Communications, and editor for the IEEE Wireless Communications Letters. He also serves on multiple TPCs in related conferences, including ACM Mobicom, IEEE GLOBECOM/ICC/GlobalSIP, and AAAI Conference on Artificial Intelligence

报告摘要：

In the recent years, deep learning has developed a lot, and showed its powerful ability. With the application of it, great progress has been made in the computer vision, speech recognition, natural language processing, the game of go, autonomous vehicles, and so on. Inspired by those advances, many researchers have restudied the wireless communication problems by using deep learning techniques. However, none of them focuses on the signal detection of MIMO system with quantized output. In this talk, we propose a deep learning estimator for the signal detection of SIMO system with 1-bit quantized output. In order to learn the detection rules, we employ a deep complex neural network(DCNN). For simplicity, we only focus on SIMO system and explore the performance of deep learning estimator compared to the traditional LMMSE and MMSE estimators. The performance of our deep learning estimator is affected by the hyperparameters greatly. Therefore, we explore the impacts of some hyperparameters to achieve the excellent performance. In addition, some important tricks, such as Xavier initialization and batch normalization, are considered to enhance the generalization performance as well.