This talk would cover the basic idea behind the encoder-decoder architecture. I would start by developing intuition about the underlying principles and would then show how to train encoder-decoders for different datasets (across domains). This would continue into a discussion about advanced applications like natural language translation, conversational agents and so.
The outline of the talk is available at : https://github.com/navagarw/PyDataTalk/blob/master/notes.md
In this talk, I would cover the following aspects:
Attendees are expected to have a basic idea of Python.
I would shortly add the url to the slide deck and code for the demo.