I keep in mind sitting in a lecture on pure language processing (NLP) with neural networks. The nice teacher confirmed us graphs on the parameter counts of the newest NLP fashions: They ranged to billions of parameters!
He then mentioned: Oh, the graph’s already outdated. Fashions are bigger now.
Wow! What a time to witness such developments; from tens of millions to billions to … trillions of parameters. Someplace round that point, I obtained hooked on all factor deep studying, doing all out there lessons on AI, plus some on-line programs (Coursera, MIT Deep Studying lectures).
The extra I realized about machine studying and deep studying, the extra I spotted: there’ll at all times be extra to find. 6 years later, I’d know barely greater than my previous machine learning-self, however I nonetheless am an absolute newbie.