Breaking down the AI suitcase.


AI can be daunting, it is important to simplify and how you break it down is vital.

A bitesize breakdown of Ai main components

According to Marvin Minksy, Ai is a ‘suitcase’ term, meaning it can incorporate so many things, and by breaking this down it is easier to understand what the general phrase ‘Ai’ is made of.

The next paragraph or two is just a very headline overview of what Ai is – needed so it can be understood.

Narrow versus General Intelligence

Firstly, when people speak of Ai it is what’s known as narrow Ai, meaning an individual task to be completed. A self driving car is an example of multiple narrow Ai systems working together.  As humans, we exhibit general intelligence, where cognitive thought intakes multiple variables and our cognition interprets these variables and provides specific reactions – human nature…and all its idiosyncrasies.

Ai includes in its basic forms

Machine Learning – the most common form of Ai is the ability of a program to see patterns in a set of variables and alter a process according to desired parameters (we will leave the deep learning and neural networks to the engineers) – google advert pop-ups reacting to your latest search…

Planning and Search – another common everyday use.  A starting point, an end point, add in some variables and create a competition – google maps directing you to your next meeting..choose shortest route or quickest route.

Natural Language Processing and generation; NLP and NLG – the ability for a ‘program’ to interpret written or spoken language input or deliver output – ask Siri to tell you a joke…

Visual Recognition – the ability of a system to identify items from images, which seems easier than it is, as systems only ‘see’ in 2D, and are only now embarking on understanding 3D – Pinterest can search for products from a picture…

Robots – are not truly part of the Ai stable, however are spoken about in the same breath.  A robot is a physical execution of a program that interacts with its environment.  A robot doesn’t have to be Ai driven and Ai doesn’t need a robot, however the two work well together – Amazon Echo and Alexa

Internet of Things (IoT) – similar to robotics, this is not classed as Ai, but is used with Ai applications.  IoT is the use of remote connection to manage items or to collect many sensor status – your ‘connected’ coffee maker turning on when you are 5 minutes away from home or your Fitbit device.

For more information see other relevant posts

Link to 

  1. SME and Ai is a perfect fit – busting 6 myths about Ai in SME
  2. How SME can use Ai to beat competition

Footnote

There are ongoing questions regarding Ai ethicacy, morality and oversight, and those questions need to be addressed today, before the conditions are hardwired into our daily lives. But until Skynet takes over and we humans battle the machines, let us use them to help us be more than we can be without them, and SME are perfectly placed to receive more value from Ai than many others ….and if any of you are thinking, I am not a bot.

After studying Artificial Intelligence with MIT, I can recommend for all those interested a free online ‘course’ offered by University of Helsinki called Elements of AI which can be found at https://www.elementsofai.com

Add Comment