Artificial Intelligence (AI) is a broad branch of computer science; but a set of technologies within the field of computing one can call learning technologies as the goal of AI is to learning new information through experience (non-symbolic-AI) or through exposure (symbolic-AI). The goal of AI is to create systems that can function intelligently and independently. In this post will do my best to provide a high level perspective on AI to give the not so familiar a decent grasp of the whole scope of the field.
AI can be divided into three functional branches:
Assisted Intelligence: widely available today, such as IPhones’ Siri and OK Google which are used improve what people and organizations are already doing.
Augmented Intelligence: emerging today, this type of AI is usually embedded enabling people, systems and organizations to do things faster. For example Netflix, YouTube, Twitter and also Google search engine AI embedded in their code to improve search results, recommendations and make your online experience more personalized and responsive.
Autonomous Intelligence: this is what has gotten all the hype and what is being developed completely autonomous self learning machines and expert systems that will out perform humans at a collection of tasks. Examples include the Narrow AI such as Alpha GO and Zero that ight now. They are called Narrow AI because they are really only able to do one thing. While us humans can do many things, as such we are pretty general learners. When a single AI can do all what we can do then we call this type of AI General AI. This type of AI is like 30 years away.
The Imitation Game
The easiest way to think about AI is in the context of a human cognition, after all humans are arguably the most intelligent creatures that we know of.
speak no evil see no evil hear no evil
We can speak and listen to as well as communicate through language. In AI development this is the field of AI is called speech recognition. Much of speech recognition is statistically based hence it’s called statistical learning. Humans can write and read text in a language. In AI development this is the field of natural language processing. We process information making sense of the world and what we see with our eyes. In AI development this is the field of computer vision. Computer vision falls under the symbolic way for computers to process information. Recently however another way to enable computer vision has been developed which we will come to later on but I think it’s worth noting now. We also recognize the scene us through their eyes creating images of that world this field of image processing which even though is not directly related to AI is required for computer vision. Humans can understand their environment and move around fluidly about their environment. In AI development this is the field of robotics.
Humans have the ability to see patterns as a result we can group like objects together. In AI development this is the field of pattern recognition Machines are much better at pattern recognition compared to us. This is because they can use more data and process more dimensions of data. In AI development this is the field of machine learning
AI family tree with symbolic learning on the left and Machine leaning on the right
The digital brain human + AI
The human brain is a network of neurons which we use these to learn. If we can artificially model and replicate some of the structures and the functions of the human brain, then may be able to also replicate some of the brains cognitive capabilities in machines. In AI development this is the field of neural networks. These artificial networks become more complex and deeper as we use them to learn and solve complex problems. In AI development this is the field of deep learning.
Dense deep neural network on right side
There are different types of deep learning and machines which are essentially different techniques to replicate what the human brain does. If we get a network to scan images from left to right top to bottom it’s a convolution neural network or CNN. A CNN is used to recognize objects in a scene this is how computer vision fits in and object recognition is accomplished.
Humans can remember the past like what you had for dinner last night. We can get a neural network to remember a limited past using a recurrent neural network
AI works can again be broken down in to two main categories; one is symbolic based and the another is data based or non-symbolic. The database side is called machine learning we need to feed the Machine lots of data before it can learn for example if you had lots of data for sales versus advertising spend you can plot that data to see some kind of a pattern. If the machine can learn this pattern then it can make predictions based on what it has learned. While humans can understand and learn 1-3 dimensions of plotted data, it’s much easier for machines. Machines can learn in many more dimensions hundred or even thousands. Machines can look at lots of high dimensional data and determine patterns from thousands of data points. Once it learns these patterns it can make predictions that humans can’t even come close to. We can use all these machine learning techniques to do one of two things:
For example, when you use some information about customers to assign new customers into a group like young and old, male and female adults then you are classifying the customer. If you use data to predict if a customer is likely to defect to a competitor then you’re making a prediction.
We have to train these non-symbolic learning machines. If you train an with data that also contains the answer (labeled data) then this is called supervised learning. For example, when you train a machine to recognize your friends by name you’ll need to identify them for the computer. If you train an algorithm with data where you want the machine to figure out the patterns then it’s called unsupervised learning; for example, you might want to feed the data about celestial objects in the universe and expect the machine to come up with patterns in that data by itself.
If you give any algorithm a goal and expect the machine through trial-and-error to achieve that goal then it’s called reinforcement learning; for example, a robot attempting to climb over the wall until it succeeds.
Hopefully this summary has left you feeling more informed about AI and it’s broad applications as you decide to make your journey towards a better understanding AI and it’s potential impacts on society.
Hope this helps…