mercoledì 25 gennaio 2017

Machine Learning For Developers

machine learning (1)

Machine Learning and its cousin Deep Learning are becoming increasingly important in a data-driven world.
The last couple of years has really seen Machine Learning (ML) take off in popularity and if you are a software developer you might be considering getting into it. After all it’s just another programming technique that you can add to your list of skills isn’t it? Well yes and no. It’s definitely not just a programming technique.
With ML there are a large number of things you need to learn. That’s true of many computing activities of course, but ML is more of a field. If you were learning to program relational databases then there’s also a lot to learn- different types of servers, how to install a database server, how to configure it, how to have your program talk to the database and of course how to write and run SQL. With ML there’s way more to learn than that.
With ML, you have fewer programming techniques to learn, it’s mostly non-programming, in particular statistics and some related mathematics. ML is a lot like applying statistical methods and that’s the hard bit to learn.

But It’s still AI?

Old school AI was about writing programs that could do things like play chess better than humans. It was very narrowly focused and became something of a dead-end. Hardware helped it win rather than “thinking like a human” which was the real objective of AI.
The field of AI went through some good times for example producing expert systems but also a couple of bad times known as Ai Winters in the 70s and 80s. It’s only since the turn of the century that ML has led to some significant progress. AI now really means Machine Learning though the term Data Science is probably more accurate.

Nessun commento:

Posta un commento