The paper is aimed to present a comparative study of different autoencoder architectures. The introduction section provides the background of autoencoder architectures and display the motivation for conducting this study. In the second part, I would like to systematically introduce the method I used to design and implement the experiment.
This section begins with data preparation, where I describe the procedures for data acquisition and normalization to ensure the dataset’s suitability for analysis. After that, I have roughly described the implement logics for each autoencoder architecture. At last, the visualization techniques are used to display the results and it can help to emphasize how these graphical representations contribute to understanding the experiment’s outcomes. In the last part, I would like to make some deepen analysis on these three autoencoder types based on the images we got on the methodology part. After that, I will discuss the advantages and disadvantages of each autoencoders, which can help us choose the most reliable autoencoder according to different dataset and conditions.
