A presentation at You Gotta Love Front End in in Nahsholim, Israel by Asim Hussain
The Future of Machine Learning & JavaScript YGLF 2019 @jawache
Asim Hussain @jawache codecraft.tv microsoft.com
https://aka.ms/jawache-cda @jawache
https://www.palinternship.com/ @jawache
@jawache
This is @EleanorHaproff’s slide Machine Learning Asim Web Development
@jawache
TheMojifier™ @jawache
themojifier.com @jawache
How to Calculate Emotion? @jawache
(1) Detect Facial Features @jawache
https://towardsdatascience.com/facial-keypoints-detection-deep-learning-737547f73515
(2) Use a Neural Network @jawache
Axon Neural Networks Dendrites Body Axons @jawache
Neural Networks 23 -0 .5 Activation Function 8.6 1 1 . 2 @jawache
Neural Networks 23 x -0.5 = -11.5 } 8.6 x 2.1 = 18.06 7.01 !-> activation(…) !-> 1 @jawache
Neural Networks 1 Output 0 0 Input @jawache
Neural Networks 4 1.1 4.2 0.3 12 3 93 @jawache
Neural Networks 4 1.1 4.2 0.3 12 8 3 - 8 = -5 93 @jawache
Neural Networks 4 1.1 4.2 0.3 12 8 3 - 8 = -5 93 @jawache
Neural Networks 4 0.1 9.2 0.2 12 8 8 93 @jawache
https://azure.microsoft.com/services/cognitive-services/face/ @jawache
https:!//<region>.api.cognitive.microsoft.com/face/v1.0/detect { “url”: “<path-to-image>” } @jawache
@jawache
Summary @jawache
Summary • Neural Networks are incredibly powerful • Conceptually, they are simple to understand @jawache
TensorFlow, MobileNet & I’m fine @jawache
@jawache
@jawache
@jawache
TensorFlow.js @jawache
TensorFlow.js Train models Load pre-trained models @jawache
MobileNet https://github.com/tensorflow/tfjs-models @jawache
https://azure.microsoft.com/services/cognitive-services/computer-vision/ @jawache
https://codepen.io/sdras/full/jawPGa/ @jawache
https://twitter.com/ollee/status/930303340516216832 @jawache
https://twitter.com/FrontendNE/status/930120267992616960 @jawache
https://twitter.com/chrispiecom/status/930407801402347520 @jawache
Summary @jawache
Summary • TensorFlow.js doesn’t have any dependancies • MobileNet is a simple way to analyse images • Azure Computer Vision API ❤ @jawache
Image2Image @jawache
DEMO https://zaidalyafeai.github.io/pix2pix/cats.html @jawache
❌ ✅ Generator Discriminator @jawache
✅ ❌ Generator Discriminator @jawache
✅ ✅ Generator Discriminator @jawache
@jawache
@jawache
@jawache
https://github.com/NVIDIA/vid2vid @jawache
https://github.com/NVIDIA/vid2vid @jawache
https://github.com/NVIDIA/vid2vid @jawache
https://github.com/hanzhanggit/StackGAN @jawache
Summary @jawache
Summary • GANs learn to generate new images • They take a lot of compute to train • But the generator model can be run in the browser @jawache
aka.ms/mojifier @jawache
Asim Hussain @jawache codecraft.tv microsoft.com
for free. You
can too.