Artificial Intelligence (AI)#
Artificial Intelligence (AI) is a technology that enables machines and computers to perform tasks that typically require human intelligence. It contains multiple subdivisions.
Categories#
Machine Learning (ML) learns from and makes decisions based on data, using algorithms to identify patterns.
Generative AI creates new content (text, images, etc) based on what it’s learned instead of just recognizing it.
Natural Language Processing (NLP) understands and interacts with human languages in a way that feels natural.
Expert Systems use predefined “if-then” rules programmed by human experts to make informed decisions.
Development#
There are a lot of steps and techniques involved in creating an ML. For more information, see:
Resources#
Some of the resources available to me to learn about/practice AI.
GeeksForGeeks#
Has a bunch of articles with and without code examples. I’ve linked a ton of them throughout my notes already.
Kaggle#
Widely known for hosting public coding competitions and datasets, they have a lot of potential in guiding growth and getting real practice
Learn section with courses like these. The tutorials don’t rely on videos, and the Exercises launch Jupyter notebook environments which you appear to be able to use without an account, and let you download the data sets. Intro to Machine Learning, Intermediate Machine Learning, Feature Engineering, Intro to Deep Learning, Natural Language Processing Guide.
Competitions also exists, and may be the optimal way for me to learn - hand me dataset and a challenge and force the actual decision-making process. Might be best to start on the Getting Started or Playground Competitions. You need to accept the rules for a competition before downloading the dataset, though you don’t have to join it. Their Getting started with NLP for absolute beginners walks you through each step of that process.
API can be pip-installed in order to programatically download/upload Datasets, interface with Competitions/Kernels/Models, etc. (see Documentation)
Machine Learning Mastery#
Spenser strongly suggested this site,
Has guides like 4 Types of Classification Tasks in Machine Learning.
Usually the guides very briefly discuss the theory and some network options, then show a code block and a graphic. -
May not be as useful given my learning style, since I need to transform information in order to store it mentally - copying and pasting doesn’t do it for me.
3Blue1Brown#
Has a YouTube Channel where he goes over a lot of complex topics, using advanced visuals to visualize advanced math.
Created the manim Python Library to programmatically animate math (documentation here).
DataCamp#
Free for me through work, has Career Tracks like Machine Learning Scientist in Python which comprise multiple courses like Supervised Learning with scikit-learn, which has chapters on Classification, Regression, Fine-Tuning your Model, and Preprocessing and Pipelines.
Courses consist of videos followed by code exams.
They’re good videos, though I don’t like that they force me out of my preferred pace, and the audio requirement makes me less likely to actually start up a course.
The code exams have you write code for a chunk of the problem at the time, which is good for learning, but it can be hard to cobble together the resulting programs and datasets into useable code for my notes.
Portfolio#
Machine Learning Mastery mentions Building an AI Portfolio, and even suggests projects in articles like 7 LLM Projects to Boost Your Machine Learning Portfolio
Kaggle’s Competitions would be a great way to go for these as well.