Making VR more interesting with JS and Brain.js
summary
Builds a web VR search-and-find game with A-Frame primitives and a Node/Express backend that uses Brain.js to train a tiny neural net (configured with three hidden layers). Training rows combine the player’s x,y,z location plus found (0/1) and steps to predict "will finish" and the server returns new item coordinates; items move closer for struggling players and farther for strong players. Code shows HTML scene setup, asset use, custom attributes and a single prediction endpoint.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:56Talk structure
- 2:19Virtual reality background
- 4:29VR app with A-Frame JS
- 17:13Adding ML to VR
- 22:55Brain.js in VR predictions
- 29:59VR and ML considerations
- 32:46Key takeaways
related talks
32mMaking VR More Interesting With React and Brain.js
25mMaking VR More Interesting With JavaScript and Machine Learning
29mWeb based Virtual Reality
53mVirtual reality for web developers
16mKeren Kenzi @ ReactNext '23 | Hello (Virtual) World
34mVirtual Reality with A-Frame
26m“There is no spoon” - Building VR experiences in the Browser (With JS)
30mGetting started with WebVR
