Artificial Intelligence. It’s a phrase that’s been thrown around a lot in the past few years, but what really is it? For most people, what initially comes to mind is probably the idea of talking robots that are capable of passing as human. For some people, AI might even spark fear - probably because of sci-fi movies like “I, Robot” that involve dystopian worlds in which the human race has been destroyed by the very robots they created to help them! In reality, AI is indeed the idea of trying to make computers think like humans, but at it’s core all it really is, is the the combination of massive amount of data with some clever algorithms.
Where did it all start?
John McCarthy in 2008
Source : Jeff Kubina
The idea of AI has been around for centuries - people have been obsessed with trying to create machines to do things for us. However, the term Artificial Intelligence was first coined by John McCarthy (an American Computer Scientist) in 1955 and shortly after in 1956 he organised the infamous Dartmouth Conference, which is where AI was first established as an emerging field of research.
Getting into the Jargon
“Machine Learning”, “Neural Networks”, “Deep Learning” - these phrases (and many more) appear time after time whenever we hear about Artificial Intelligence. They’re all important constituents which make up the field of AI, but what do they actually mean?
Machine Learning
Machine learning can broadly be split into two categories: supervised learning and unsupervised learning. Take, for example, trying to predict housing prices based on how many bedrooms a house has. We would give the computer a training set - this would contain the number of bedrooms for lots of different houses, and how much that house sold for. We use this data to train the computer, and from it the computer learns how to predict housing prices of other houses based on the number of bedrooms. This is supervised learning because we are giving the computer examples and telling it the outcome in order to train it. In contrast, in unsupervised learning, data is given to the machine but there are no outputs (e.g. the price of the house) - instead we ask the computer to make inferences based on the inputs only.Neural Network
Neural networks are algorithms which have been designed to mimic how the brain works. They are a type of supervised learning algorithm. We give the computer input data with a labelled response, and we design some sort of path in which the data is processed - this path is the neural network. The paths in between the input and output layers are called hidden layers. Algorithms can have any number of hidden layers - this is decided by the programmer.Deep Learning
Deep learning is any type of machine learning that involves neural networks, but these neural networks have lots and lots of hidden layers (hence the term “deep learning”.Computer Vision
Computer vision is anything concerned with teaching computers how to “see”. Essentially, we want computers to draw data from images and videos, and this is generally done using deep learning algorithms.Natural Language Processing
Natural Language Processing (or NLP for short) is an interdisciplinary field which draws on knowledge from linguistics, computer science and AI. It involves training computers how to process human language - whether as speech or text - and often make inferences from this data.
Source: DataCamp
AI in the Modern World
It goes without saying we, as humans, are unbelievably dependent on technology everyday. What you might not have realized, however, is that you probably use AI all the time. How did you find this article? Maybe you asked Siri to find you an article about AI, or you typed “Ariticial Intelligence” into Google and this article popped up. Either way, AI was at the core of your search. Another example that we all take for granted are spam filters in our email inbox - gone are the days of missing an important bill because it was hidden in a cloud of spam! Instead, our inboxes automatically send irrelevant emails to the spam folder (most of the time…).
But there are some even more interesting uses of AI that researchers are working towards in the hope of improving the world we live in:
Improving accuracy and wait time of cancer Diagnoses
Reducing bias in criminal justice
Driver-less Cars
Understanding and fighting climate change
Intrigued?
Although it sounds super complicated, there are some amazing resources out there to teach you more about the algorithms behind AI and the impacts AI can have on society. Here at Educademy, we offer courses in AI for kids as young as 7 in two different programming languages - Scratch and Python.
If you want to read some more about AI in the modern world, I would highly recommend reading Hannah Fry’s Hello World: How to be Human in the Age of the Machine. It’s a great introduction for pretty much anyone of any age or ability into the world of AI. She also hosted a podcast with AI Giant DeepMind in which she interviews some of the best researchers in the field. You can listen to the podcast here or on whatever app you use to listen to podcasts.
Cover Photo Source: Franck V
Author: Richa Lad