As the last example of this chapter, we'll add text description of the problem into observations of our model. We've already mentioned that some problems contain vital information given in a text description, like the index of tabs needed to be clicked or list of entries that the agent needs to check. The same information is shown on the top of the image observation, but pixels is not always the best representation of a simple text.

To take this text into account, we need to extend our model's input from an image only to an image and text data. We have worked with text in the previous chapter, so a Recurrent Neural Network (RNN) is quite an obvious choice (maybe not the best for such a toy problem but it is flexible and scalable). We are not going to cover this example in detail but will just focus on the most important points of the implementation (the whole code is in Chapter13/wob_click_mm_train.py). In comparison to our clicker model, text...