Digital Transformation Today

What is Artificial Intelligence (AI)?

You’ve seen it mentioned in articles, press releases, and in conference speeches – and for good reason. No longer the stuff of science fiction, AI.jpg Artificial Intelligence (AI) has become one of the most talked about topics in business and beyond. For all the excitement surrounding AI, details tend to be a little thin. A company or product is “Powered by Artificial Intelligence” or “Uses Advanced AI to Solve <your problem goes here>.” But what does this mean? What are they really doing? What is AI?

When John McCarthy coined the term in the 1950’s, AI was “the science and engineering of making intelligent machines.” Technology has advanced since McCarthy’s definition, but the premise still stands. Francois Chollet, an AI researcher at Google, more recently summed up the entire field of Artificial Intelligence as “the effort to automate intellectual tasks normally performed by humans.”

Simply put, in today’s world “Artificial Intelligence” refers to systems capable of automating tasks that would normally require human levels of intelligence.

Early AI systems relied on extensive libraries of hard-coded rules written by experts on the problem they were attempting to solve. A classic example would be a chess AI composed of if-else conditional logic rules written by an experienced chess player. This type of AI, known as Symbolic AI or Expert Systems, was popular in the 80’s and 90’s, but have been over taken by modern approaches based on machine learning.

Instead of relying on rules written by experts, modern AI systems utilize Machine Learning to extract information from data then make optimal decisions towards achieving their goal. That goal could be winning a game of Go, recommending a TV show, or identifying objects in an image.

How Does AI Use Machine Learning?

Think of an old Expert System as the process of combining a set of data and a set of rules as inputs to produce answers. Compare this to Machine Learning which takes data and answers as inputs to learn the relationships hidden in your data, which can then be used on new data to produce answers.

AI blog diagram

Modern AI systems use Machine Learning (ML) algorithms to understand the data they are being fed and make decisions based on the algorithm’s output. In practice, this means AI systems are highly specialized due to the nature of how Machine Learning algorithms are trained and used. The same AI that makes smart investing decisions will not be able to identify faces in a photo or recommend a great restaurant. Each of those AI utilize ML systems that were purpose-built for their task.

The narrowly focused systems we use today are known as Weak AI in comparison to other types of AI such as Strong AI (conscious or sentient machines –SkyNet) and Artificial General Intelligence/AGI (AI that performs across any problem, but not sentient). Though currently the stuff of fiction and philosophy, Strong AI and AGI are both heavily pursued at research institutions around the world.

AI, ML and Deep Learning

Another buzzword you may have seen around AI is “Deep Learning.” Deep Learning refers to a category of Machine Learning algorithms known collectively as Artificial Neural Networks which are trained on large sets of data to find even the subtlest relationships to use towards successfully completing their task. Advances in Deep Learning are partially responsible for the explosion of AI and machine learning driven systems in technology today. Not all AI systems are built on Machine Mearning, and not all Machine Learning based AI systems use Deep Learning (see diagram).

ai diagram

The world of AI and Machine Learning is vast and constantly changing. New algorithms, architectures, and systems are constantly being released, pushing the state of the art further and further. With an understanding of how AI and Machine Learning see the world and make decisions, you will be better prepared to harness their benefit for yourself and your company.

If you’re interested in how machine learning and artificial intelligence can benefit your organization, or if you have in-depth questions about a specific project, schedule a consultation online, or give us a call at 240-406-9960.

Previous Post

Next Post