I am yet to attend the talk. But after quick investigation, this is what I understood!
Attached is a high level quick component flow diagram, explaining overall steps.
I am not sure if I am understanding your question correctly or not.
Aliases and entity linking appears to be manually curation.
For example:
{"entity_id":"a1","name":"Machine learning","description":"Machine learning (ML) is the scientific study of algorithms and statistical models..."}
{"entity_id":"a2","name":"Meta Language","description":"ML (\"Meta Language\") is a general-purpose functional programming language. It has roots in Lisp, and has been characterized as \"Lisp with types\"."}The entities a1 and a2 share a common alias called ML
Disambiguation is done at Linking Process layer using the context.