Digital Library[ Search Result ]
Capsule Neural Networks as Noise Stabilizer for Time Series Data
Soyeon Kim, Jihyeon Seong, Hyunkyung Han, Jaesik Choi
http://doi.org/10.5626/JOK.2024.51.8.678
A Capsule is a vector-wise representation formed by multiple neurons that encodes conceptual information about an object, such as angle, position, and size. Capsule Neural Network (CapsNet) learns to be viewpoint invariant using these capsules. This property makes CapsNet more resilient to noisy data compared to traditional Convolutional Neural Networks (CNNs). The Dynamic-Routing Capsule Neural Network (DR-CapsNet) uses an affine matrix and dynamic routing mechanism to train the capsules. In this paper, we propose that DR-CapsNet has the potential to act as a noise stabilizer in time series sensor data that have high sensitivity and significant noise in real world. To demonstrate the robustness of DR-CapsNet as a stabilizer, we conduct manual and adversarial attacks on an electrocardiogram (ECG) dataset. Our study provides empirical evidence that CapsNet effectively functions as a noise stabilizer and highlights its potential in addressing the challenges of preprocessing noisy measurements in time series analysis.
Visualization of Convolutional Neural Networks for Time Series Input Data
http://doi.org/10.5626/JOK.2020.47.5.445
Globally, the use of artificial intelligence (AI) applications has increased in a variety of industries from manufacturing, to health care to the financial sector. As a result, there is a growing interest in explainable artificial intelligence (XAI), which can provide explanations of what happens inside AI. Unlike previous work using image data, we visualize hidden nodes for a time series. To interpret which patterns of a node make more effective model decisions, we propose a method of arranging nodes in a hidden layer. The hidden nodes sorted by weight matrix values show which patterns significantly affected the classification. Visualizing hidden nodes explains a process inside the deep learning model, as well as enables the users to improve their understanding of time series data.
Search
Journal of KIISE
- ISSN : 2383-630X(Print)
- ISSN : 2383-6296(Electronic)
- KCI Accredited Journal
Editorial Office
- Tel. +82-2-588-9240
- Fax. +82-2-521-1352
- E-mail. chwoo@kiise.or.kr