How to get started with AI, explained to 13-year-old…
Muhammad Anas
Exploring the advent of AI | Founder SnazzyPlum | Author at Machine Learning Pills
Whenever you start doing something new, the first question that pops into your mind is, where should I begin? And if recently you started to venture into AI, you may have stumbled upon this question. You may have even searched on it, but to be honest, it’s a complicated answer to find. So let’s find that answer once and for all.
First let’s clear out what is AI, ML, and DL — it’s simple there are a subset of one another. ML is a subset of AI and DL of ML. AI is the bigger picture and things keep going smaller as you keep diving in — as AI has no limit, it’s your brain which your limit. AI is broad and might look all very overwhelming, however, I’ll try to clear all your doubts by the end of your read.
I’ll be mainly focusing on Machine Learning in this blog.
Where do I begin?
First of all, we need to choose a programming language. There are tons to choose from, but I recommend Python as it is both beginner-friendly and most of the AI/ML applications in the world are made in Python programming language. Tesla’s self-driving AI and Netflix’s recommendation systems — are all built-in pure python. So it’s an all in all, start with Python and thank me later.
Just to note, Python doesn’t limit you to only Machine Learning and AI, you could do so much more like creating the backend of websites or automating some boring, repetitive task. All can be done with Python’s vast count of libraries for all sorts of tasks.
Done with Python. What Next?
Now we need to move on to the libraries that we will be using to do Machine Learning.
For ML, we mainly end up using Sciket-Learn. Sciket-Learn mainly focuses on giving us pre-maid algorithms so we don’t have to code these long and difficult algorithms ourselves which saves us a lot of TIME. You could train an ML model with Sciket-Learn in a few minutes, but if you had to code those lengthy algorithms yourself, you’d go mad, at least I would.
BONUS: We also use other libraries like Numpy, Pandas, and Maplotlib to help us do Exploratory Data Analysis so we can decide what algorithms to use and where to use them. Filter the useful data and discard the useless data. Exploratory Data Analysis is an important part of any Machine Learning project. But what do these 3 libraries do?
If you are looking for sources to get to know these Libraries, then?Corey Schafer?and?Keith Galli?are worth watching, they both give hands-on ways to learn about these 3 libraries which I think is amazing!
You forgot about the Math…
Nope, I didn’t. What to do about the extensive amount of math included? If I’m bad at math does it mean I won’t be able to do ML? NOT AT ALL, it’s jus=t a myth knowing all that math. Let’s break this down.
领英推荐
If you’re a teenager, maybe around 13–16 years of age you probably don’t know anything about Linear Algebra (Vectors and Matrices) or Calculus (Derivations and Integrations). And if you haven’t, truly speaking, you don’t need to either — at least not as a teenager. These concepts are hard to grasp, just focus on your school math and I think it’s more than enough.
But if you’re in college, try to learn these topics — at least Linear Algebra. Khan Academy is a great place to start. But if you think you don’t need to, then you don’t need it much. You can make ground-breaking projects without knowing a bit of the math behind, get a $200k job without the math, but learn to stand out. And I think to do so, knowing the basic math is a must!
But lemme say it again, MATHEMATICS FOR ML IS NOT IMPORTANT, at some points (1 in 100) maths might be helpful, otherwise, you never tend to use much. So if you want to simply stand out and understand the logic behind how the Machine Learning algorithms work then knowing some math doesn't hurt. And anyways, Linear Algebra is easy, it’s Calculus that takes dedication so if you’re a teen wanting to know some maths behind an ML model, then Linear Algebra should do the work.
Books or Courses…
As said for Python and Data Science you begin with the?Data Science Course, there is this amazing course by IBM on Cognitive AI (Both on Python and Data Science). It is one of those courses, which provide a certificate of accomplishment for free. It also has a?Machine Learning course, which is undoubtedly one of the best!
Note:?You may see people recommending,?Machine Learning by Andrew Ng?hosted on Coursera?— it’s not a bad course. It’s the BEST, and that is why you should take it. It had been there since the beginning of time. The Holy Grail of Machine Learning Courses. Something to note here, in June 2022, Andrew is releasing his updated version of this course as it has been 10 years old, and much of the content is outdated.
If you’re more likely to read books then the book?Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems?by Geron Aurelien is an absolute must. It covers extensive topics in ML with a lot of code and hands-on practice.
Then you could go on and learn Deep Learning via Pytorch or Tensorflow or Keras for many different tasks, like Natural Langauge Processing(VOICE TO TEXT) and Image Recognition, kinda like the Siri on the iPhone and Google Assistant on Android. And maybe one day you could have Jarvis in your bedroom!
A quick side note…
Machine Learning is overwhelming, it’s a black box with so many things to learn. And you’ll fail and make many mistakes, and the only way to excel in the domain of AI/ML is to learn from your mistakes. Make projects and make mistakes and learn! The learning of Machine Learning never ends. Have a great learning journey!
Conclusion…
I expect you got your answer to your long-awaited question. And you finally found out where to start and how ML learning never ends due to the endless possibilities. Your mind is your limit. And do note, learn from your mistakes!