Join the Revolution
summary
An open source indecent‑content checker NSFW JS gets trained for weeks on a home machine and runs entirely in the browser using TensorFlow.js (examples also use brain.js and TensorFlow Lite); lightweight models range from ~48 KB to ~23 MB, a browser‑trained rock‑paper‑scissors model reaches ~93% validation accuracy, and the talk cites a melanoma study trained on 129,000 images matching dermatologists plus Microsoft’s machine vision that labeled 125 million buildings — useful but prone to false positives and ethical tradeoffs.
AI-written summary · about
chapters (10)
- 0:00Start
- 0:29Gant Laborde
- 2:17INFINITE RED
- 2:27Machine Learning?
- 6:43Machine Learning What does it do?
- 13:32What should it do?
- 18:04Machine Learning Applications
- 18:19What is it?
- 21:02From Where?
- 44:11Review
related talks
54mAdventures in Javascript Machine Learning
29mSuperpowers for next gen web apps: Machine learning
54mDeep Learning in JS
15mMachine learning for next gen web apps with TensorFlow.js
39mMachine Learning Magic for Your JavaScript Application
37mJavaScript beats cancer
29mImage Processing in the Browser
36mPractical Machine Learning for Front End Developers