There are a lot of people talking about AI, machine learning and deep learning right now. And unfortunately it seems like only half of them know what they talk about and don’t only use the words because it’s a trendy topic.
If you are lost in all the buzz-words (like I was a few months ago) and are not so much into reading (like me), here’s a very short introduction and a list of good videos to get started and learn what these topics are all about.
AI can be understood as a parent-category covering all efforts to solve very computational-hard problems in an intelligent way and to make machines act more intelligent (like humans). It covers topics like search (e.g. find the fastest route between to cities), planning (to make a robot navigate and reach a given goal) and machine learning. AI if often viewed relatively to the current state of computer intelligence and computational power. Things that were seen as intelligent behaviour a few years back are now taken as given and are not “intelligent” anymore.
In machine learning, algorithms learn from existing labeled data to solve problems and it’s not required to hard-code these rules in the algorithm, e.g. to build a system that can say if there’s a dog or a cat in a given image, the algorithm is shown lots of pictures of cats and dogs to learn how to distinguish them by himself.
Deep learning is a part of machine learning, where a neural networks, that is a multi-layer system with a many neurons similar to the structure of the human brain, is trained to solve very complex machine learning tasks (like the dogs/cats image example above) with a lot of data. This topic is so hot right now because now for the first time we have the cheap computational resources (mainly GPU’s) and lots of data to train the models with.
AI, Deep Learning, and Machine Learning: A Primer (Andreessen Horowitz)
A Friendly Introduction to Machine Learning (Udacity)
Artificial Intelligence is the New Electricity (Andrew Ng)
There are more interesting videos videos by Andrew Ng on youtube and he also published a free machine learning course on coursera (but that one is very technical/theoretical).
Artificial Intelligence, Revealed (Facebook):
For all videos (including the more detailed technical explanations), check out the official facebook page.
Deep learning simplified:
For all videos, check out the series on youtube.
Build your own
When you want to learn more about how to build neural nets yourself:
- Learn the Python programming language (e.g. on codecademy, codeschool or udacity)
- Take a look into Googles TensorFlow framework as well as keras for easier model-implementation and go through the mnist tutorial (to get a feeling how machine learning looks like) and take a look at this video:
3. Take the free Udacity Deep Learning Course by Google
Happy learning :-)