Part one: The first class
After 6+ years of study of a machine, both in the study and industry, is very interesting to see how far the field is. I still remember that I sitting in the Memina rooms, listening to all things ml: Deeply learning, random forests, natural networks, environmental consideration, …
A specific NLP talk represents my memory, where we discussed the fastest development in the field. We discussed Vanilla's attention on last week and now they looked for the Technique for a Transformer. The Great Teacher showed US graphs with models of models. Then we look forward to the latest progress, and we became clear: Any image is already within the month. At that time when my ML journey was completely starting, and it is already in a new published condition and again insane.
Since then, my trip continued in several stages. From an exchange of ML people, it turns out that they have the same experience: From the beginning to a fanatic seller, the trip can be separated from the following three phases:
Category I: We started (~ 1 year; this article)
SECTION II: Intermediate (~ 1 to 3; next)
Phase III: Advanced (~ 5 + years; coming)
