Welcome to Neural Notes

During my 10,000 hours of late-night tinkering, studying, and building AI models, I've learned one immutable truth: the more you know, the more you realize you don’t know. From debugging sentiment analysis models to fine-tuning derivatives prediction systems, and even teaching computer vision networks to "see," every milestone has taught me not only about AI but about learning itself.

Throughout these years, I’ve had many opportunities to explain concepts to people at different levels of expertise. Whether it was a casual conversation or an in-depth technical session, one thing became clear—I didn’t know as much as I thought I did until I tried to teach it. That realization drives this newsletter.

Neural Notes serves a dual purpose:

1. To identify and close my own knowledge gaps while creating a resource for other practitioners, hobbyists, and the curious alike.
2. To dispel the misconceptions about AI that run rampant in the age of viral news, overblown claims, and AI fear-mongering.

In an era where generative AI dominates headlines and terms like "machine learning" have entered dinner table conversations, it's more important than ever to separate fact from fiction. There’s no shortage of hot takes online, ranging from predictions of AI utopia to existential doom. Neural Notes aims to ground these discussions in clarity and context.

What You Can Expect

I’ve outlined a year's worth of content (subject to change as feedback and new ideas arise), We’ll start with foundations and core model intuition. Included in each article will be a code playground to mess around with and get as "tangible" and "tactile" an experience as is possible with virtual things. Here's the breakdown for Quarter 1:

Quarter 1: Foundations of AI and Model Intuition (Weeks 1-13)

Goal: Build foundational knowledge and break down key concepts for readers of all levels.

Week 1: Welcome (this issue)

Week 2: What is AI, Really?

• We’ll explore the differences between AI, machine learning, and deep learning

Week 3: The Anatomy of a Model: Input, Output, and Parameters

• A breakdown of what goes into a machine learning model, explaining it step-by-step.

Week 4: Soft Intro to Data Science with Python

• We’ll cover the tools you need to get started: from importing data with Pandas and Numpy, to visualizing trends using Matplotlib. (and maybe bashing tensorflow a little bit for fun;)

Week 5: Intro to Neural Networks

• Understanding the structure of a neural network and how "neurons" make predictions.

Week 6: Activation Functions 101

• Why activation functions are the secret sauce that gives neural networks the power to solve non-linear problems.

Week 7: Cost Functions and Loss: Why Your Model "Learns"

• How models quantify their mistakes and learn from them through the optimization process.

Week 8: Gradient Descent: AI's Learning Path

• How neural networks update their parameters through the famous gradient descent algorithm.

Week 9: A Beginner's Guide to Data Preprocessing

• Prepping your dataset by cleaning, normalizing, and encoding it properly—because garbage in, garbage out!

Week 10: Overfitting and Underfitting: The Balance of Learning

• Understanding the dangers of memorization vs. under-generalization and how to strike the perfect balance.

Week 11: Understanding Model Metrics: Accuracy Isn't Everything

• We’ll dive into key evaluation metrics like precision, recall, and the F1-score to measure model performance more holistically.

Week 12: The Dataset Dilemma: How Data Shapes AI's Behavior

• The ethical and practical implications of the datasets we feed into our models.

Week 13: Recap with Hands-on Demo

• A small end-to-quarter project: building a simple, functioning regression model in Python to reinforce everything we’ve covered.

Real-Life Application and AI Incubator Updates

As part of this newsletter, I’ll also share my journey building models for my AI incubator, Eleanor AI. This incubator focuses on practical applications of AI across different domains, and you’ll get an inside look at the challenges, experiments, and breakthroughs along the way. Whether it’s the thrill of discovering a better model architecture or the frustration of debugging data drift issues, you’ll be right there with me for the ride.

Why Eleanor AI? "The future is here-it just isn't evenly distributed." At Eleanor AI, we believe that advanced technology should be accessible and affordable for everyone, not just a select few. Our goal is to build innovative, affordable AI solutions that address real-world problems and democratize the benefits of AI.

How You Can Contribute

I want this to be a two-way conversation. I’m always eager to hear your feedback, thoughts, and burning questions about AI. Are you curious about a concept, a new research paper, or a headline? Let me know!

Neural Notes is as much yours as it is mine—a place where we can dive deep into the fascinating and sometimes perplexing world of AI.

What’s next? In Week 2, we’ll dig into "What is AI, Really?" and clarify what distinguishes AI, machine learning, and deep learning. (We'll even get into a very interesting concept...Markov Chains!)

Thanks for joining me, and welcome aboard!