The Future of Machine Learning & JavaScript
summary
Built a working Moji Fire Slack bot that detects faces and returns emotion scores using Azure Face API (attributes like anger happiness scored 0–1), experiments run fully in-browser with TensorFlow.js in-browser via WebGL and preloadable models, MobileNet recognizes ~1,000 classes so accuracy is often low on niche objects, and pix2pix/GAN image-to-image demos generate plausible fills after heavy offline training then export small models that run client-side.
AI-written summary · about
chapters (25)
- 0:00Introduction
- 1:56Udemy
- 3:15AI JavaScript Rocks
- 3:46What is JavaScript
- 4:55mojifire
- 5:19MO defies it
- 6:12MO on Slack
- 7:13Emotion
- 7:40Facial Features
- 8:48What are Neural Networks
- 9:34How to Code
- 10:24Activation Function
- 11:38Network
- 15:24Public Datasets
- 16:03Common Problems
- 16:41Azure Face API
- 17:59Summary
- 18:33Demo
- 19:57Tensorflow
- 20:56Tensorflow JS
- 22:20MobileNet
- 25:07Computer Vision
- 28:39Summarize
- 29:38Image
- 38:34Conclusion
