How AI Works on Mr. Pointing

How AI Works - Chapter 5

Sun, 28 Dec 2025 00:00:00 +0000

Convolutional Neural Networks: AI Learns to See

While the last chapter toted how advanced neural networks were, they weren’t always able to select the correct features, determining the right amount of dimensions, along with their inability to learn, lead to our next topic; Convolutional Neural Networks. What they do differently than the networks we looked at in the last chapter is they actually generate new representations of their inputs while also being able to classify them. This is called end-to-end learning.

How AI Works - Chapter 4

Mon, 27 Oct 2025 00:00:00 +0000

Neural Networks: Brain-Like AI

We must understand going forward that even though we’ll liken the brain’s neurons to that of connectionist neural networks, they are not the same. Similar, but operate on a totally different level. At a basic level, brain neurons are in an off state until turned on. Artificial neurons are the exact opposite; also containing inputs and outputs but continuously calculating some mathematical function.

A good example is that of a light switch versus a dimming switch. The light switch turns on and off, while the dimming switch varies between different states of being on and producing light. These are the biological and artificial in essence, with some slight overlap that we don’t need to concern ourselves with.

How AI Works - Chapter 3

Tue, 30 Sep 2025 00:00:00 +0000

Classical Models: Old-School Machine Learning

Obviously, we need to start with baby steps. No one picks up a new hobby and starts with the most advanced material. We’re going to look at 3 of what we call the classical models; neither symbolic nor connectionist AI models that aren’t as advanced as any of the neural networks we’ll take a peek at in chapter 4.

Nearest Neighbor

This model is so simple, that the training data is the model, meaning there is no training. If you get an unknown input, you classify it to what it’s closest too, and that’s that. Regardless, they’re still super useful, and are a good representation of actual real life data.

How AI Works - Chapter 2

Wed, 10 Sep 2025 00:00:00 +0000

Why Now? History of AI

Like we learned in the last chapter, AI has been around for a while. In the grand scheme of things, this “while” is seen as a blip, still in it’s infancy. It certainly did not just appear with ChatGPT. Let’s take a look at the landscape that got us here.

Types of AI

I know I said we’d get a lot of definitions out of the way in the last chapter, but I lied. We have some more to go. AI can fall into two major categories; Symbolic AI and Connectionism. Where symbolic AI takes advantage of visual symbols and logical statements to build it’s “intelligence”, connectionism will try to build it’s intelligence using a compilation of simpler, smaller components. These are significant because it’s how our human brains work. We consider symbolic to be “Top-Down” (high level tasks getting broken up) and connectionism as “Bottom-up” (smaller tasks combined to create larger complex tasks).

How AI Works - Chapter 1

Tue, 09 Sep 2025 00:00:00 +0000

And Away We Go: An AI Overview

Consider this the master chapter for all basic AI definitions. Everything following this chapter will build off of the points brought up in this chapter. Some foundational stuff right here.

Again, we should start at the most basic; Artificial Intelligence is the process of training a machine or computer to act like an intelligent human being. This original terminology was first used by John McCarthy in the 1950’s!