Training vs. Inference
Interactive lab demystifying AI training vs. inference with zero black boxes.
Episodes & demos
- S01E01 Playhttps://tc.livive.org/ ► Watch (opens in a new tab)
About
Training vs. Inference is a static single-page app that demystifies how AI actually works for non-technical learners. It walks through a multi-module curriculum covering hardware and memory (storage, RAM, inference), a 1-neuron binary classifier, a 3-neuron positive/negative/neutral spectrum, and a multi-task enterprise router, plus a sentiment-classification demo. Built with a zero-black-box philosophy: every weight, bias, tally, and calculation is rendered directly in the DOM in human-readable numbers. Implemented in vanilla HTML5, CSS3, and ES6+ JavaScript with zero external dependencies or build tools, and hosted on GitHub Pages.
Summary
A browser-based interactive learning lab that teaches the difference between untrained inference, training, and trained inference through inspectable, hand-verifiable numbers.
Tech
- HTML5
- CSS3
- JavaScript
- GitHub Pages
Tags
- ai
- education
- interactive
- machine-learning
- neural-networks
- sentiment-analysis