Artificial Intelligence, Enough of the hype! What is it?

[Mattab]

There is a lot of hype around Artificial Intelligence (AI) and as I mentioned in my last blog there is weak and strong AI. Personally I prefer the term narrow instead of weak for AI as the AI we have today is focused on narrow or very specific tasks. AI in of itself is a large field and the best way of getting to understand AI is to get a grounding in the basics. I like to think of AI as being an onion, as it has layers of depth starting on the outside with AI and with Deep Learning at the centre. Don’t worry I will explain what these terms mean in a little bit.

Before we dive into some specifics I want to explore a couple of concepts, the first one is that AI is something that can be done ‘out of the box’, unfortunately it cannot. Narrow AI is brittle. What do I mean by that, well take your AI powered personal digital assistant on your phone, do you ever ask a question that it just does not understand? Secondly, narrow AI can fail with unpredictable results, there are a number of well documented cases where Narrow AI has caused issues. That said AI has had some notable successes, I mean EA Sports predicted the 2018 World Cup Winners. One of the great things about the field is that it is learning and evolving all the time, which means it is constantly getting better and better.

Let’s get to some terminology, sorry for the boring bit, we will avoid the complex mathematics (comments below if you want me to write something about it in a future blog), think of AI like owning a car, you do not need to know how the engine works to drive it. If we take AI as a field then it is a machine that is capable of learning and performing tasks to human levels. This is the stuff of science fiction, we are not there (yet) and there is much debate among the academic community around how long that this will take. So the next level is go and create a machine that learns. What do I mean by that, well traditional computing requires that a programme is developed which contains all the rules that are needed in order for the programme to make a decision. Namely if this and this, then that. These rules can get pretty entangled, especially when the volume of data increases and gets more complex. Think about writing a set of rules that would enable a computer to recognize a human being in the same way we can. One that could differentiate from apes and other animals and also recognize the difference between real images rather than cartoon characters or painted images, whilst also recognizing that tattoos, jewelry and clothing etc do not define the underlying human. Something that comes naturally to us would require many many rules to describe for a computer. Imagine if you missed one? More on the ethics of AI in a later blog.

Anyway back to rules, a computer programme based on static rules is not going to be very good at predicting an outcome, more on that in a bit. There are a number of ways that a machine can be artificially intelligent but the most successful is Artificial Neural Networks. Inspired by biological brains these use artificial neurons to form the basic computational unit and networks are used to describe the interconnection between these artificial neurons. Networks which use multi-tier or multi-stage neural networking for feature extraction are called deep learning. So Deep learning is a popular way to achieve machine learning.What-is-Artificial-Intelligence-The-Layers-Explained.jpg

To think about it in a slightly different way AI covers anything which enables computers to behave more like human brains, but without the emotion. Again, this is the stuff of science fiction and a long way from what we can achieve today. So Machine Learning is the subset of AI that deals with the extraction of patterns from data sets. This means that the machine can find rules for optimal behaviour but also can adapt to changes in the world, it can learn. What this means is machines can find rules in subject matter that manual creation of the rules is either extremely costly or practically impossible.

There are a number of ways of achieving Machine Learning such as Random Forests, Support Vector Machines or Naïve Bayes Classifiers, which I will explore in a future article. However, due to advances in technology the most successful has been models inspired by nature, which created a wonderful entity in the form of the biological brain. Artificial Neural Networks are statistical learning models that are derived from the ones nature created and Deep Learning is a specific class of Machine Learning algorithms (aka mathematical equations) which use complex neural networks. These mathematical models are capable of predicting an outcome based on a complex set of input data.

To give you an example if I am playing the card game Texas Hold ’Em, with an incomplete set of data I need to work out or even better predict if the cards in my hand will win me money. Playing games of chance when you do not know what your opponents has requires the Artificial Intelligence to make a guess. The best mechanism of making a guess with incomplete data is to use a Deep Learning algorithm which is fantastic at spotting patterns and making informed guesses or to more precise a prediction.

AI is a very broad field, with a huge depth of possible applications and the number of use cases are only limited by your imagination. HPE has enormous experience in finding the right use case for you and putting in the right platform in place that scales to meet you AI needs. Think of AI as a complex mathematical equation that can digest vast amounts of data and create rules without having them explicitly programmed. AI will change the world. The possibilities are very exciting, as AI is helping tackle some of the world’s most pressing issues (i.e. world hunger, antibiotic resistant viruses & energy) but it is also helping with problems business are facing every day. AI will help us understand the world around us and gain greater insight into either our business or our planet. We are generating so much information that it has already become impossible for us to sift through it for knowledge. Information is being bombarded at us at an ever increasing rate, we need help and AI is the answer.