Thread by @markawest

There’s a lot of confusion about Artificial Intelligence and how it relates to Machine Learning and Deep Learning. Heres how I explain these terms to my customers.

Thread.

(2/6) Artificial Intelligence is a general term, encompassing everything to do with making computers more like us. At this level there is as much focus on philosophy (what is intelligence) and psychology (how do does the human mind work) as there is on technology.

(3/6) Machine Learning is a specialisation of AI, which focuses upon the use of applied statistics to build build predictive and descriptive models, which focus on finding correlations between historical data and a desired outcome. Example methods include SVM and Random Forests.

(4/6) Deep Learning is a specialisation of ML & focuses on the use of bio-inspired neural networks (such as Convolutional Neural Networks). DL is often better than simple ML techniques for dealing with large feature sets, but DL models give less insight to their inner reasoning.

(5/6) Many go straight for Deep Learning when a simpler Machine Learning solution would suffice. As with all tools, the most important thing is to understand how to use them. This is where Data Science and knowledge of maths and statistics comes in.

(6/6) Getting good results from Machine Learning / Deep Learning depends on having plenty of good quality data. Bias in your data will result in a Biased model. Google "Amazon CV" for an example of AI Bias.

Bonus Tweet: I write more about this and related factors at bouvet.no/bouvet-deler/6….

Or you can watch this presentation: jfokus.se/jfokus19/talks….

Happy Sunday!