AI learning journey step 1 : First steps into AI and resources to get started

This is the first article of a series of articles about my learning journey into AI. I hope it will be useful to anyone.

Feeling lost in all the content

In our current world, where we do not make difference anymore between AI generated content and real content, I felt a little lost when starting to learn this new subject. Social media and youtube were all full with tutorials, opinions, and courses about AI, and all this content was not organized in a way that I could easily grasp. So I decided to go back to the basics, and how I started learning the many tech topics in the past. I spent time checking the resources, looking into online communities and forums, filtering to only keep the most reputable ones.

I also decided to document the journey, as a way to keep track of what I am learning, as well as an accountability tool to keep me motivated.

I am not an expert, I am just sharing what I am learning at the moment.

What do we mean by an AI model ?

• An AI model, is, in very basic terms, just a mathematical function that takes some input and produces some output. Nothing more. It's main purpose is to make predictions based on the input data you give it.
• The model start with a random set of parameters, and then it uses a set of data called the training data to adjust those parameters in order to reduce reduce errors in its predictions. This process is called training the model.

Small explanation of the buzz words : Machine learning, AI, Deep learning, LLMs, etc...

• AI is a field of computer science that develops systems capable of performing tasks that typically require human intelligence, such as learning from experience, recognizing patterns, and making decisions.
• Machine learning is a domain in computer science that focuses on algorithm that can make predictions or decisions based on data, and without explicit programming.
• Deep learning is a subset of machine learning that, tries to extract patterns from data using multiple layers that we call neural networks.
• LLMs, or large language models, are a type of deep learning model that is trained on a large corpus of text data.

How the model learns ?

To make predictions, there are different strategies, and each depends on the type of data you have, and the task you want to do. There is 2 main big categories :

• Supervised learning: The data given to the model has both inputs and expected outputs, and the model learns to map the input to the output. As simple example of this category can be, a simple house price prediction model that uses house size (input) to predict price (output). In a very simple form, this can be expressed just as y = ax + b.
• Unsupervised learning: The data given to the model, has only the input. The model has to find, and construct patterns and structure in the data in order to be able to make predictions. A common example of unsupervised learning is customer segmentation. A model can try, based on a large dataset of customer purchase histories, to analyze the data to discover hidden patterns and automatically groups customers with similar behaviors together. The model can do this without any labels of predifined customer categories.

What is an AI agent ?

An AI agent, is just a model, that can take actions based on a brain, a set of tools, and the external environment. In details :

• The Brain (Language Model): The brain of the AI agent. It does the thinking, analyzing, and planning.
• The Body (Tools): The interfaces, APIs, and functions that the language model can choose to execute. Tools bridge the gap between abstract thought and concrete execution.
• The External World: The environment outside the AI agent, that the agent should connect to to do some action, such as public apis, external tools like google maps, slack, ...

Resources ?

With the many resources there are out there, I felt a little lost at first. Here are some of the resources I found useful to start with :

• Datacamp : It is a great platform that has many courses on machine learning and AI. They have tracks that are well organized and that can help you learn step by step.
• Deeplearning.ai : It is a site made by the famous Andrew Ng, and it has many courses on machine learning and deep learning. It can be a little bit too theoretical. What I am doing is, try to follow the courses while also practising the concepts.
• Machine learning for everybody : This is a great video done by Kylie Ying in collaboration with freeCodeCamp. It is very beginner friendly, and it gives a good overview of the main concepts in machine learning.
• MIT Introduction to deep learning: This is a playlist of videos from MIT about main concepts to know in deep learning. I found the 2 professors are very good at explaining the concepts.