Using JavaScript to Teach Machines Basic Shit Things
summary
Claims that you can do real machine learning in JavaScript, showing a Naive‑Bayes linear classifier that sorts tweets into four categories via the Twitter API, a case where it mislabels “women in tech” as unrelated, and a JS neural network that actually generates tweets from the collected dataset and lives on GitHub. Notes tradeoffs: JS works and integrates with web apps but often reinvents wheels better served by Python and can be slower for big data.
AI-written summary · about
chapters (10)
- 0:00Machine learning with JS
- 1:28Machine learning and AI
- 2:30Supervised vs unsupervised learning
- 4:36Smarter Twitter bot
- 6:23Bias-variance trade-off
- 8:51Discrete vs continuous data
- 9:35Deep learning overview
- 10:40Convolutional vs artificial neural networks
- 11:42Can machine learning be done in JavaScript
- 12:47JS libraries and implementation
