Perhaps one of the most unique facets to Artificial Intelligence is the advent of Machine Learning, which is essentially defined as the ability for a computer to learn without being explicitly programmed. Now, how would we do that, you would ask.
Well, to start off, Machine Learning has a few components. First of all, we need to train the computer by supplying the inputs and the associated outputs for data. Now, this can be done in multiple ways. For example, on way is to do this with image recognition. Here, we would take an input file of images with the associated tag. One real-life usage of this is for dog vs. cat classification. We could easily obtain a sample size of dog and cat images from the internet, then we would have to classify each one, and eventually upload all of these so that our machine could learn to recognize the images.
Each time that our computer, or machine, goes through the images, it will also check to see if it got the right answer, or result, that we obtained. Thus, after a mass number of pictures have been run, our new model could predict whether an image has a cat or a dog, when given simply test images.
What makes Machine Learning so powerful is that the images could be entirely different than the ones that were used in the training process. Thus, similar to facial recognition, since your face does change, your phone can consistently make sure that the face is correct.
I hope this blog post was helpful for learning the basic systematic process of machine learning. We will be diving into much more specific details in the coming weeks. See you next time AI Lovers!