AI and JavaScript, no You're Not Dreaming
summary
Demos run live in the browser and on Node using TensorFlow.js: GPU-accelerated WebGL inference and training, model conversion from Keras, and transfer learning. Examples include body-pix real‑time body segmentation for background blur, speech-commands driving a canvas game, and a retrained action classifier built from YouTube frames with image augmentation. Models export as JSON served over HTTP, training can run on CPU/GPU/TPU in Node, and most prototypes take minutes rather than weeks.
AI-written summary · about
chapters (12)
- 0:00Introduction to AI in web
- 2:31Why use AI in browser
- 5:46TensorFlow.js overview
- 9:27Browser demo: BodyPix
- 13:39Live camera body segmentation
- 15:53Accessibility with AI
- 16:35AI in Airbnb's workflow
- 18:00Speech commands demo
- 21:26Creative AI projects
- 25:00Node.js and server-side AI
- 27:37Retraining models in Node
- 30:26Future of AI in web
related talks
41mAI in Web: TensorFlow js
18mHow to AI in JavaScript
46mArtificial Intelligence with JavaScript
34mA Powerful Introduction to Machine Learning with Js
44mArtificial Intelligence and Machine Learning in the browser
39mMachine Learning Magic for Your JavaScript Application
24mTensorFlow.js + Node
42mAI in React JS