Biosignal Generation and Latent Variable Analysis With Recurrent Generative Adversarial Networks (2019)
Open in webOpen in zoteroOpen pdf
1 Abstract
The effectiveness of biosignal generation and data augmentation with biosignal generative models based on generative adversarial networks (GANs), which are a type of deep learning technique, was demonstrated in our previous paper. GAN-based generative models only learn the projection between a random distribution as input data and the distribution of training data. Therefore, the relationship between input and generated data is unclear, and the characteristics of the data generated from this model cannot be controlled. This study proposes a method for generating time-series data based on GANs and explores their ability to generate biosignals with certain classes and characteristics. Moreover, in the proposed method, latent variables are analyzed using canonical correlation analysis (CCA) to represent the relationship between input and generated data as canonical loadings. Using these loadings, we can control the characteristics of the data generated by the proposed method. The influence of class labels on generated data is analyzed by feeding the data interpolated between two class labels into the generator of the proposed GANs. The CCA of the latent variables is shown to be an effective method of controlling the generated data characteristics. We are able to model the distribution of the time-series data without requiring domain-dependent knowledge using the proposed method. Furthermore, it is possible to control the characteristics of these data by analyzing the model trained using the proposed method. To the best of our knowledge, this work is the first to generate biosignals using GANs while controlling the characteristics of the generated data.
2 NOTES
In this paper the authors investigate the quality and diversity of data generated using GANs (with LSTM-based architectures). On the matter of classification, they did test it on the three chosen datasets and observed that by increasing the data (by a lot) always reduced the error rate, but it was no the focus of the work. There are two main things in here: the use of Dynamic Time Warping (DTW) to compare similarities between samples and use of Canonical Correlation Analysis (CCA) to control the latent space from the GAN. The first one is clear on its usage. The second is quite interesting, since they could not simple pass the whole signal to control, they instead chose to set some features: maximum value, the point of maximum value, the minimum value, the point of minimum value, maximum-to-minimum interval length, mean amplitude, and the mean frequency. This led to some interesting results, changing the signal quite a lot but without losing its main structure.