Slide 1

Slide 1 text

Machine Learning Manuela Rink, Software Engineer, Microsoft for the noob!

Slide 2

Slide 2 text

No content

Slide 3

Slide 3 text

What… “Recognize handwritten text numbers in an native iOS app … offline!” … could possibly be so hard?

Slide 4

Slide 4 text

What do I know about ML?

Slide 5

Slide 5 text

Python Machine Learning basic knowledge again Python convert data for model usage some other things – dunno yet? use CoreML to predict correct results I’ll just need a bit of…

Slide 6

Slide 6 text

Quick tip from a colleague who is totally into ML: “Just use an SVM with the MNIST digits database to get the model. And don’t forget to tweak it with a decent grid search!”

Slide 7

Slide 7 text

No content

Slide 8

Slide 8 text

Where do I even get started?

Slide 9

Slide 9 text

Step 1 “Embrace being the noob – and just run the code”

Slide 10

Slide 10 text

Step 2 “Create and train your model - then convert to CoreML”

Slide 11

Slide 11 text

Step 3 “Understand what you’ve just created”

Slide 12

Slide 12 text

Step 4 “Get your input in shape, RLY!”

Slide 13

Slide 13 text

Prep’ing your data for predictions

Slide 14

Slide 14 text

Step 5 “Make your prediction – ALL THE RESULTS!”

Slide 15

Slide 15 text

No content

Slide 16

Slide 16 text

It’s dangerous to go alone… https://github.com/codePrincess/doodlingRecognition +

Slide 17

Slide 17 text

http://scikit- learn.org/stable/auto_examples/classification/plot_digits_classification.html#sphx-glr-auto- examples-classification-plot-digits-classification-py http://yann.lecun.com/exdb/mnist/ https://developer.apple.com/documentation/coreml/converting_trained_models_to_core_ml ?language=objc https://www.gitbook.com/book/leonardoaraujosantos/artificial-inteligence/details https://docs.microsoft.com/en-us/azure/machine-learning/studio/algorithm-cheat-sheet … and this

Slide 18

Slide 18 text

Merci :) Manu Rink Software Engineer [email protected] @codeprincess says

Slide 19

Slide 19 text

[1] Scientific Droid - https://techfinancials.co.za/2017/08/08/future-artificial-intelligence/ [2] Chappie - http://www.newsweek.com/artificial-intelligence-scientists-racist-sexist-robots-ai-693440 [3] Machine thinking - https://www.sciencenews.org/article/machines-think-predicts-future-artificial-intelligence [4] Headless Human - http://www.ttec.com/sites/default/files/styles/article_main/public/perspectives1.jpg?itok=o4K9k0zs [5] Learning robot - http://robohub.org/wp-content/uploads/2017/01/machine-learning.jpg [6] ML Workbench - https://images.anandtech.com/doci/12508/azure-machine-learning-studio-predictive-score-experiment.png [7] Zelda Sword - http://piq.codeus.net/picture/143569/wooden_sword_the_legend_of_zelda_nes_ [8] Embrance - https://memegenerator.net/instance/51017062/oso-hormiguero-thug-embrace-the-change [9] HeMan - https://www.storegrowers.com/google-analytics-for-ecommerce/ [10] Results - http://www.providentmediagroup.com/blog/testimonials-study-results-and-gift-card-winner/