Hello To Machine Learning
Cutting the long ML story short.
After demystifying AI, it’s now time to turn our attention to Machine Learning. Let us clear all the confusion in our minds about machine Learning. (Click here for the previous blog)
Myths of Machine Learning
I need to be a computer science student to understand Machine Learning….
No, not at all. You needn’t be. You weren’t a civil engineer while you fit a nail in the wall. Nor were you a mechanic while you drove your car. Machine Learning is an art and it has to be appreciated. You aren’t a football expert but you do appreciate when someone scores a goal, right? Same is with Machine Learning, you needn’t be a statistician or mathematician or a computer science graduate to comprehend Machine Learning.
You can be a silent spectator in AI (just like an audience) or maybe a die-hard fan of AI (some fans go really crazy) or perhaps an expert in AI. To be any of these you need to be able to appreciate what the computer does and how it does. Remember that we taught the computer to learn, so it isn’t hard to learn how the computer does it.
I am weak in Maths and Statistics, I don’t like to solve equations, hence I feel ML is not for me.
You don’t have to get scared. Neither do you have to solve equations or do lot of calculations. Most of the part is already implemented in Python and its packages. You have to just use the tools. You need not know how the engine of the car works but still, you can drive really well right?. Though expertise in Math helps to get an in-depth understanding. It is not always required. Just 12th standard Math is more than enough, along with very basic statistics.
You can just say that “Okay, the Math part will be done by Python. Let me do the part after it”. Machine Learning uses math because the computer cannot understand our language. The computer can understand only equations and numbers. But this doesn’t mean that you need to only know equations or math. You can stay 50 feet away from math and still learn Machine Learning.
Well if you don’t believe the above line then as you continue in my future blogs you will.
Do I need to code to do Machine Learning or all is done by the computer itself?
Computers are not smart enough to do “Machine Learning” by themselves. We need to code to teach the computer. The programming part is not the conventional way in which you might have done programming. It is simpler and more straightforward. Most of the part is implemented already in Python. You just have to learn how to implement the “models”. It isn’t that difficult but yes, you have to code. Coding is the only way in which we can teach the machine, isn’t it?
I don’t know to code. How do I learn ML?
You do not require very high level of programming. Basic Python knowledge is enough. In case you are a non-programmer. You can still read my blogs very comfortably; I will separate the coding section and help you understand and appreciate what and how Machine Learning works. Don’t worry if you cannot program, you will be alright. Eventually, you will fall in love with ML, (through my blogs of course) learn to code and then code to learn again.
I don’t have a good computer, also I don’t know how to set up my computer to do Machine Learning. How do I code then?
Relax you needn’t set up your local machine to code. You could just use free cloud platforms such as Google Colab or Kaggle Kernels. These contain all the packages that you require to code for free. They run in your web browser and require no setup. We will use them soon.
What should I code?
We will get there soon probably after a few blogs.
What is Machine Learning? …. Truth in simple words.
What does Machine Learning exactly try to do?
Machine Learning tries to learn a relationship between the input data and the output. We train “models” which for this job. In simple terms, the job of a model is to take input data, learn a function say “f” which is f(input_data) and give us the desired output.
The input data maybe text data, images or even videos. The output is something that we expect from the model. E.g. We do weather forecasting in our minds (We do this every-time before a cricket match 😃). We take into consideration the humidity, clouds, location, past experience, month, etc. These are our input attributes. The final output is “Yes” if it rains and “No” if it doesn’t.
The same task now we want to automate. That is we want a “model” to take these inputs do computations as we do it in our head then tell if it rains or not.
What are these “models” how do I make them ?
The most important thing what we study in ML is these various “models” that map the input to the output. You don’t have to create a new model necessarily. There are tons of models that people have researched for past more than 30 years. You need to learn them and understand when which model works, how does it work and when which doesn’t and why it doesn’t.
You need to use these “models” to model the given data. Sounds very simple right? I will teach you all the models then you try them one by one and YAY!! you know Machine Learning. The bitter truth: this isn’t that easy as learning models takes time.
How do the “models” learn to map the input data and the output ?
No model can learn on its own, we need to teach it. The soul of Machine Learning lies in teaching the models to learn. We will discuss this in the next blog.
The Art of Machine Learning.
Machine Learning is all about mapping the input data to the output. To make this possible we use the existing models, sometimes we also create our own models.
The art of Machine Learning lies in mastering the way to analyze the data and interpret it efficiently. Models are like paintbrushes and colors you have to decorate the beautiful data given to you.
Art needs to be appreciated (not all can master art, but everyone can appreciate) the same is with AI. If you are a common man trying to understand what AI is, remember that AI is the ability of the computer to interpret the data as humans do.
That’s what Machine Learning is all about. It is pure art which all of us can understand. The art lies in making the machine learn. The art is to develop mechanisms for machine to learn.
In the next blog, I will teach you this art of learning. Stay Tuned 😃
The next blog is released 😃 click here .
About Me: -
I am a University student trying to bring AI closer to the common man. Through my blogs I want to educate everyone about Machine Learning. I intend to reach out everyone, even if you have no clue what AI is.
You can view me on linked by clicking here.
Thanks to Anushchandra Shetty for the Proof-reading and suitable edits.
