This is the most exciting part of the word embedding experiment. Seeing is believing, and visualizing the embedded space should be able to provide us with an intuitive understanding of what embedding results look like. The TensorFlow community has provided a tool that we can use to visualize the embedded space. This is called Embedding Projector:
With this tool, we can create visual information about the embedded space in a 2D or 3D manner. The algorithms we can use here are PCA and t-SNE. t-SNE is another dimensionality reduction algorithm that is quickly gaining popularity in the deep learning community.
We can ...
