Keynote talk - The future of machine learning & JS
summary
Argues that JavaScript can run and consume real ML now: uses TensorFlow.js to run models in browser or phone, uses Custom Vision to train image classifiers by uploading tags (minimum five images per tag) and calling the prediction API, and wires Azure Text Sentiment into a Logic Apps workflow that posts 98% positive scores to Slack. Demonstrates a doppelgänger app that returns similarity percentages (often wrong, skewed toward Arnold Schwarzenegger, e.g. 31% mismatches). Notes training can need thousands of images (a farmer used 7,000 and waited days).
AI-written summary · about
chapters (6)
related talks
24mThe Future of Machine Learning & Javascript
15mThe future of machine learning using JavaScript
34mThe Future of JavaScript and AI
34mThe Future of JavaScript is A.I.
48mThe Future of Machine Learning & JavaScript
34mBuilding Machine Learning-powered Applications with JS
46mArtificial Intelligence with JavaScript
44mArtificial Intelligence and Machine Learning in the browser