Have you ever wondered how you as a sentient human is able to identify images? , how you are able to tell the difference between looking a table and a chair , or a cat and a dog.
What if you travel thousand of light years into space and meet aliens and make friends with them ,One day you're trying to explain the significant physical differences between a cat and a dog, how would you do it?
I guess you'd start from describing whiskers on the cat,more teeth on the dog , their relative sizes and probably the jaw shapes of both animals. Those characteristics are what you call "features".
Let's probably assume you don't have a photo like this one you'd have to explain major features to differentiate these animals
Now , away from advanced lifeforms who could actually comprehend your explanation.
Let's say you're trying to explain it to a computer,
....Can you?, We know computers understand stuff in zeros and ones so how would we be able to even give instructions in the first place 😂. Well, hold up... Computer programmers are people who writes programs to make a computer perform a specific task.
Now, let's assume that you're a computer programmer who commands a computer system. Now how can you explain the pressing issue of differentiating between a cat and a dog to a computer?Actually if you can it'll be an Herculean task having to hard-code all the pixels and their specific locations.
Moreso what if you're presented with a different breed of dogs and cats to work with, what if the location of cats and dog in the image is changed or even worst-case scenario you are presented pictures of horses and now you have to re-adjust the entire program to meet those edge cases.
Machine learning is way of finding patterns in data through feeding tons and tons of data to a computer system. From our last paragraphs problem, instead of we writing a program to differentiate these two animals , we could give the computer thousands of images of these two animals and with the help of some algorithms, find suitable patterns and differentiate to even human level capabilities between these two animals
Machine learning could be applied not to only our problem but to many other problems , it could perform not just differentiating between images but differentiating between text , automated (self) driving cars and even recently ,generating images that have never existed before.
Above are just a few of the applications of machine learning , it possesses a lot of untapped potential and has a lot in store for the future
Hey, check out this app I built to differentiate dogs from cats using neural networks , a form of machine learning.
Cats-and-Dog-Classifier
Thanks for reading 🙏
Top comments (0)