Understanding Machine Learning-1

Source: Deep Learning on Medium


Go to the profile of FikriGelgit

These article series is based on my learning process of Machine Learning. Because of this process is continuous, at that point I don’t know how many articles I will write about it.

Why am I writing about?

I believe in Feynman Learning Technique which consists of the 4 steps. This technique simply starts with the selection of concept. After studying this concept 1–3 hours, writing down in plain language what you understood from the concept, is the second step.The most important thing is that you should pretend you are teaching it to a child. After this step, review your lacking knowledge. And finally, just simplify your language as much as possible.

Second reason why I am writing?

This is some type of contest between me and my wife. I claimed that I could teach the artificial intelligence to her in 30 days at least conceptually. The main hardest point is that I have little knowledge about the Machine Learning at the beginning so I should learn firstly.

So let’s the contest begin….

Photo by Abigail Miller on Unsplash

AI? ML? DL?

The main confusion at the starting is what these three concepts are and how they are related to each other.

Start with Artificial Intelligence

The main term that covers all of three concepts mentioned above is AI ( Artificial Intelligence ). Simply, AI is a much more wider term for creating a machine that can perform tasks just like a human. There are a lot of discussions about AI whether it is good for the mankind or not. The smartest and distinguishable scientist of the last 50 years, Stephan Hawking warned the mankind about AI before he passed away

“The development of full artificial intelligence could spell the end of the human race.”

The second warning comes from the well-known billionaire, Elon Musk. He believed that AI is much more dangerous than nuclear weapons.

<blockquote class=”twitter-tweet” data-lang=”en”><p lang=”en” dir=”ltr”>Wow, if even Pinker doesn’t understand the difference between functional/narrow AI (eg. car) and general AI, when the latter *literally* has a million times more compute power and an open-ended utility function, humanity is in deep trouble</p>&mdash; Elon Musk (@elonmusk) <a href=”https://twitter.com/elonmusk/status/968560525088055296?ref_src=twsrc%5Etfw“>February 27, 2018</a></blockquote>
<script async src=”https://platform.twitter.com/widgets.js” charset=”utf-8″></script>

Photo by Evan Dennis on Unsplash

So what is Narrow AI? Is there any other AI?

Surprisingly the answer is YES.

Narrow AI is the only type of AI that humanity has achieved so far. It is good at performing a single task, such as predicting sales figures, weather forecast, playing chess. There are so many applications and systems around us which widely use AI concepts. The most well-known is self-driving car technology. It has a single task, driving a car, but it uses several Narrow AI’s.

The other one is General AI. Simply, it mimics the human. It can understand its environment and behave just like human. Also it has power to learn from their mistakes and make changes to fix them. In other words, just like Human evolution has been continuing for million years, general AI can evolve itself but in very very short terms.

There is also Super AI phase after General AI. When AI becomes much smarter than mankind in almost every field, we will enter Super AI phase.

Ok, I think it is enough explanation for AI. In the series of my learning journey, I will focus mostly on practical side of AI, both technical and business use cases. Although I also have some concerns about it, I will not dig into philosophical debates about whether it is good or not.

What about Machine Learning

Machine Learning is just a way of achieving this AI dream. We can build a system by hard coding everything theoretically but it wouldn’t be easy when the system gets bigger and bigger. Instead of hard coding, machine learning tries to train an algorithm so that it can learn by itself, how.

In the upcoming articles I will start to explain what Machine Learning is, how It can be used…

Deep Learning

Deep learning is one of many approaches to machine learning. Other approaches include decision tree learning, inductive logic programming, clustering, reinforcement learning, and Bayesian networks, among others.

Deep learning was inspired by the structure and function of the brain, namely the interconnecting of many neurons.

I will focus mainly deep learning upcoming articles after some explanation articles of other Machine Learning approaches.