Beginning Anomaly Detection Using

dilated temporal convolutional network and the encoder- 

decoder based temporal convolutional network. It is important to note that both 

models involve 

supervised anomaly detection, although the encoder-decoder TCN is 

capable of semi-supervised anomaly detection since it is an autoencoder.

 Dilated Temporal Convolutional Network

In this type of TCN, we deal with a new property known as a 

dilation. Basically, when the 

dilation factor is greater than 1, we introduce gaps in the output data that correspond to 

the dilation factor. To understand the concept of dilation better, let’s look at how it works 

for a