A Quick Introduction to the World of AI, ML, and Neural Networks
We’ve probably heard this a thousand times now but here we go again. AI is on the verge of taking over the world by storm. Cars are now able to drive themselves, robots can now learn to detect various types of cancers and diseases in humans, your cashier in your local Walmart has become a robot, and much more.
Don’t worry. There’s no need to panic. You’re not gonna lose your job or gonna see robots marching into your home tomorrow. But seriously, imagine how cool it would be to have your robot do all your chores in the house when you can just chill and watch Netflix? Well, that’s the power of Machine Learning and Artificial Intelligence. Artificial Intelligence is essentially a field of computer science that aims to help computers achieve human-style intelligence. Machine Learning is an approach that is taken to achieve this stage of human-like intelligence.
You are probably starting to wonder, “Okok I get that Machine Learning is all about programming stuff to do something but how does it really work?”
Well, here’s the answer:
By now, you’re probably ready to close this tab in disappointment. But, let me offer you something: stay on and I promise that I won’t waste any more of your time ;)
So, as you could see from the image, Machine Learning uses something known as a neural network to get this done as seen above. Of course, it’s called a neural network since Machine Learning focuses on attempting human-like intelligence. Thus, the name, “neural network.” Think of this as a human brain. Now, let’s think of those circles in this artificial brain as a neuron. These circles essentially transmit data or in this case, numerical values to other circles which also contain numerical values. From there, an operation is performed each layer of circles which continues to decrease up until we have the desired quantity of output.
In a neural network, we have 3 types of layers:
- Input Layer
- Hidden Layer
- Output Layer
Let’s say that we wanted our neural network to recognize different types of clothing. Our input layer would simply be images of clothing that are pre-labelled for our neural network to learn from. After a test image is passed into the neural network, the network tries to find correlations, or patterns, between the pre-trained data. This would be our hidden layers where all these take place. Our output layer would be the final prediction. What type of clothing is it? A shirt? A pant? A Shoe? What? We would use an activation function to help determine this. This can include functions like the sigmoid activation function or the softmax function.
Before I move on, let me just summarize everything before your head starts spinning:
- Machine Learning and Artificial Intelligence are fields of computer science that aim for computers to have human-like intelligence
- A neural network is a series of “neurons” that work together to predict something
- A neural network has 3 layers: an Input Layer to take in the input image, a hidden layer to perform operations to help predict what the given input is, and an output image which uses an activation function to state the final output.
Here’s a video and another medium article that you can take a look at for better understanding:
Understanding Neural Networks: What, How and Why?
Neural networks is one of the most powerful and widely used algorithms when it comes to the subfield of machine…
Well, thanks for reading and hope you learnt more. Stay tuned for more updates and new projects!