MACHINE LEARNING:
ML FOR THE REST OF US
Introduction
Imagine you're standing at a crowded railway station waiting for a friend. When your friend arrives you have no problem picking her out of the crowd. Recognizing people's faces is something we humans do effortlessly, but how would you program a computer to recognize a person? You could try to make a set of rules. For example, your friend has long black hair and brown eyes, but that could describe literally billions of people.The truth is that we can recognize people without ever really knowing how we do it. We cannot describe every detail of how we recognize someone. We just know how to do it.
Face recognition is an example of a task that people find very easy, but that is very very hard for computers. These tasks are often called artificial intelligence or AI.How do we get the computer to understand images, to recognize faces, or tell the difference between cats and dogs? How can we program a computer to do it if we don't know how to do it ourselves? The short answer is, we can't. But a new approach called Machine Learning is radically changing how we create software to solve these problems.Instead of programming a computer by telling it every detail of how to do a task, we teach it by giving it examples of what to do.Machine Learning uses statistical algorithms to learn from examples. We call these examples data and we say that the computer learns from data. The most popular method of moment is deep neural networks.The methods used in deep learning are basically the same as those used in the 1980s.So why if they suddenly started working a lot better?Because now we don't have better algorithms, we just have more data!
Machine Learning is about creating statistical programs called models. A model takes an input and gives you back an output.A Machine Learning algorithm takes the examples and uses them to train the model. This means that they adapt the details of the model so that it maps the inputs in the example data to the corresponding outputs. So the input can be almost anything you can represent in a computer, photographs, speech, music, web pages, bank accounts, social media profiles, DNA sequences, legal decisions, disease symptoms, news stories, astronomical data, or cute kitten videos. Almost anything you do on a computer could have Machine Learning applied to it and the basic techniques are pretty much the same.
Most machine learning algorithms are based on statistics and could be quite mathematical and complex. But to give you a sense of how some algorithms work, let’s take an example.You have lots of pictures of cats and lots of pictures of dogs. You’re given a new picture and have to decide if it’s a cat or a dog. How can you do it?One simple way of doing this is to look through all the original examples and find the one that’s most similar to the new picture and use its class. It’s a very simple method but can actually work quite well. In fact, it’s an established algorithm with its own name nearest Neighbors. It’s called that because you’re classifying a picture based on the nearest example or neighbor.One simple way of doing this is to look through all the original examples and find the one that’s most similar to the new picture and use its class. It’s a very simple method but can actually work quite well. In fact, it’s an established algorithm with its own name nearest neighbour.
Data Representation
A computer is a machine that can represent many, many different types of information. But at their most basic level, these types of information are all represented in terms of bit, the most fundamental unit of digital information.A bit is very simple yet powerful. A bit can be one or zero when it's dealing with numbers, black or white for pictures, or locked or unlocked for security features..But how can bits represent landscape paintings?A single bit can represent black or white. But by putting together lots of black and white pixels, you can make a picture.f you understand how bits work together, then you will have got to grips with a really fundamental part of how computers work.
ML in practise
Often, we humans can't do tasks like recognizing faces perfectly, so how could we expect computers to do so? But that makes it quite hard to work out if a system is working acceptably because you'll never get it to be 100% correct. A good start is to see how much of the data it classifies correctly. Once your algorithm is finished learning, you can test it by classifying each item of the training data and see what percentage it gets right.
Social impact of ML
One of the most impressive uses of machine learning on language has been speech recognition. Of the sort you can see an automatic assistance like Amazon Alexa or Apple Siri. The fact that a computer can understand your question and come up with the correct answer is pretty astounding and it's all down to machine learning on large amounts of recorded voice data. The vast amounts of data we're producing on the Internet are likely to feel new amazing breakthroughs that apply machine learning to new areas.data. You're also likely to encounter machine learning if you do online shopping. The product recommendation that sites like Amazon give you are chosen using machine learning. If Amazon's recommendation seems uncannily accurate, it's because it's learned from it's hundreds of millions of customers, and that's one of the characteristics of modern machine learning. You've come to know by now that machine learning works better the more data you have. So the companies that can use it most effectively are those with the most data.The reason a lot of people are now excited about machine learning is that it's having massive success outside of a lab in real world applications that we use all the time
So this article is directed at people from all backgrounds and all levels of technical expertise to get up and running with Machine Learning. I hope that you encounter various “aha!” moments because the conveyed content in this article was just immediately enlightening
I hope you like this article please share and also clap.
Thank you!!
