Using LSTM Autoencoders on multidimensional time-series data

In this article, I'd like to demonstrate a very useful model for understanding time series data. I've used this method for unsupervised anomaly detection, but it can be also used as an intermediate step in forecasting via dimensionality reduction (e.g. In a nutshell, this method compresses a multidimensional sequence (think a windowed time series of multiple counts, from sensors or clicks, etc) to a single vector representing this information. With an effective encoder/decoder, we can use the latent vector as an input in a multilayer perceptron or as another set of features in a larger multi-head network. I am not going to cover the details of LSTMs, or Autoencoders.