Machine Learning with Deep Neural Networks in the Browser
summary
Interactive demos show classification of colored points running entirely in the browser using TensorFlow Playground and convnetjs, training with gradient descent (examples run 20 to 400 iterations) and reporting both training and test loss to catch overfitting. A concrete puzzle requires three neurons to separate a circle from its surround, GPUs speed up the matrix multiplications, and the live demos successfully learn the decision boundary without any Python install.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:05What is Machine Learning
- 4:31Machine Learning Playground
- 7:04Neural Networks
- 8:36TensorFlow Playground
- 15:36Fully Connected
- 16:26Softmax
- 18:56Overfitting
related talks
26mMachine Learning in the Browser with Deep Neural Networks
36mMachine Learning in the Browser
35mMachine Learning in the Browser with deeplearnJS
34mMachine Learning in the browser
34mA Powerful Introduction to Machine Learning with Js
48mDeep Learning in the Browser
35mThe Future of JavaScript & Machine Learning
34mA Powerful Introduction to Machine Learning with Js