Lets Start Learning AI From Zero to Hero
If you don't know anything about AI, this article is for you.
There is no doubt that many of you are fascinated by AI technologies such as ChatGPT, Bard, Midjourney, and Leonardo.ai, among others. Of course, great minds are behind these significant technologies. However, I assure you that all of them started from “zero”.
If you are following me, you have likely read this note before. But if not, no worries; let's address this topic again. Let’s answer why you should choose AI.
Why AI?
According to the research, being done by Goldman Sachs, Market interest in AI has increased dramatically. The share was %0.81 in 2016, but in 2023 it is %16.63, which is over a 20x increase in 7 years.
That means many jobs will be replaced by AI. It's a golden age for AI enthusiasts, both now and in the future.
How to learn AI if you are a total beginner?
To learn complex concepts, you should first divide them into smaller chunks.
Divide and conquer.
Julius Caesar
To break down AI, one could start with data science, proceed to machine learning, and then delve into deep learning. However, these are multidisciplinary fields.Let's focus on data science first.
Data Science consists of multiple disciplines, such as:(in brackets, corresponding Python libraries are written.)
Data Exploration ( NumPy, Pandas, Matplotlib)
Data Manipulation (NumPy, Pandas)
Data Analysis ( NumPy, Pandas)
Data Visualization ( Matplotlib, bokeh, Plotly, seaborn)
Machine Learning ( sci-kit learn, Tensorflow, Keras )
In the beginning, all you have to do is learn how to apply those disciplines (Python libraries) above by using programming languages. And the number one option is Python when it comes to Data Science.
At the beginning, all you need to do is learn how to apply these disciplines (Python libraries) using programming languages. The number one option for data science is Python.
Why Python?
There is a lot of research available on the web that you can read. I am not familiar with Java; however, I started learning Data Science with R programming in 2017. It is safe to say that Python is way ahead of R in DS. The syntax is much easier, and the environment is better. What does that mean?
That means the people who are using Python are more active, and you can debug your code by searching through the web easily. I remember that I researched on the web to debug my code for almost 40 minutes when I was using R. But you can find an answer in Python easily.
Of course, nowadays, you can debug your code by using ChatGPT, Bard, Claude 2, or that kind of large language model easily; however, sometimes, even these tools might not help you.
Where to Start?
For total beginners, it is good to start with some basic concepts in Python. The best approach is not to rely on outdated, boring courses. With the release of powerful large language models (LLMs) such as ChatGPT, Bard, Claude 2, and others, learning has become significantly easier.
Here are Python basics;
Python Basics
Variables and Data Types
Operators
Conditional Statements (if, elif, else)
Loops (for, while)
Functions
Lists, Tuples, Dictionaries, and Sets
File I/O
Advanced Python Concepts
Object-Oriented Programming (OOP)
Error and Exception Handling
Modules and Packages
Decorators and Generators
You can learn all of them by just using the right prompts first, and doing exercises about them by using Python.
Data Exploration
At this stage, after having a Python experience, doing a lot of exercise will make you the best. While this phase is relatively easy, it is the most crucial step to understanding the data you will work with.
Basically, knowing a little bit of statistics, Python, and especially Pandas library, you will be just fine. But if you don’t know these concepts fully, you can use data exploration prompts as a shortcut at the beginning.
Data Visualization
After exploring your data, this step will be your publicity agent. It is not just discovering hidden patterns, it is like the cover page of your project to the world. That’s why your task here is crucial if you want to tell the world that this project is the one.
Machine Learning
Let’s divide Machine Learning to conquer it when it comes time.
Simple Machine Learning
Supervised Techniques
Regression
Classification
Unsupervised Techniques
Clustering
Neural Networks and Deep Learning
Computer Vision
CNN
Natural Language Processing
RNN
Reinforcement Learning
To learn all of these in Python, you should learn first how these concepts work, then practice with Python. Of course, you will find ;
Prompts
Jupyter Notebooks
Articles
about these concepts on this Substack and my notion page (for my paid subscribers), but before starting Machine Learning, you should learn a little bit of math too.
Math For Machine Learning
This is not mandatory, but if you want to be the best, you should know these concepts.
Mathematics Foundation
Basic Algebra
Statistics and Probability
Linear Algebra
Calculus
Final Thoughts
The last thing I want is to intimidate you. I was intimidated by many when I was starting out, which prolonged my learning phase. Your learning curve will depend solely on your willingness to learn and your passion.
I have seen many struggle for a long time to understand AI, but I have also seen individuals become professionals in just 5-6 months, securing their first freelance job within 3-4 months of starting.
Your progress will largely depend on your eagerness to learn and your passion. It's reassuring to know that learning is easier now than ever in this age of AI. Thanks for reading!
“Machine learning is the last invention that humanity will ever need to make.”
Nick Bostrom