En introduktion till maskininlärning och djupinlärning

One example of machine learning is Spotify, which deploys an algorithm in order to get to know your musical selections and use this data to predict which other bands and artists may be of interest to you. Spanning multiple industry sectors, machine learning automates basic as well as more complex tasks.

Deep learning is a subset of machine learning. While a machine learning model must know how it is supposed to make an exact prediction, deep learning can be compared with how a person would perceive something, think about it and draw a conclusion. In order to achieve this, deep learning uses a layered structure of algorithms know as neural networks. Its design is inspired by the biological neural network which is used by the human brain. Google’s AlphaGo is a good example of how deep learning is used. Google created a computer program which taught itself to play Go, the abstract board game, which is famous for requiring considerable human intuition. Utilising deep learning, it learned how to play at a professional standard by competing against other professional Go players instead of learning when to perform a specific move, as is the case in a standard machine learning model. Employing a model based on neural networks, AlphaGo was able to defeat the 2016 Go world champion, which was considered to be out of the question for technology based on algorithms, that is, traditional machine learning. AlphaGo has been hailed as a milestone in the development of artificial intelligence.