2020/03/03 by Hao Ge, Ge, Hao, Xiaoguang Tu +7
Computer Science · Engineering · Mathematics · #68T45 #Adversarial Robustness in Machine Learning #Anomaly Detection Techniques and Applications #Artificial intelligence #Artificial neural network #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Convolutional neural network #Data mining #Digital Media Forensic Detection #Enhanced Data Rates for GSM Evolution #Epistemology #FOS: Computer and information sciences #FOS: Electrical engineering #Field (mathematics) #I.4.m #Interpretability #Interpretation (philosophy) #Machine learning #Mathematics #Mechanism (biology) #Rationality #Relation (database) #Signal Processing (eess.SP) #acm:68T45 #cs.CV #eess.SP #electronic engineering #information engineering #msc:68T45
paper · pdf · doi:10.48550/arxiv.2003.01413
published in arXiv (Cornell University) (Cornell University) · Deep learning, adversarial example, interpretability
arxiv created 2020/03/03 · openalex publication_date 2020/03/03 · arxiv updated 2020/03/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
The interpretability of Convolutional Neural Networks (CNNs) is an important topic in the field of computer vision. In recent years, works in this field generally adopt a mature model to reveal the internal mechanism of CNNs, helping to understand CNNs thoroughly. In this paper, we argue the working mechanism of CNNs can be revealed through a totally different interpretation, by comparing the communication systems and CNNs. This paper successfully obtained the corresponding relationship between the modules of the two, and verified the rationality of the corresponding relationship with experiments. Finally, through the analysis of some cutting-edge research on neural networks, we find the inherent relation between these two tasks can be of help in explaining these researches reasonably, as well as helping us discover the correct research direction of neural networks.