![NEURAL NETWORKS
A Quick Introduction to Neural Networks
https://ujjwalkarn.me/2016/08/09/quick-intro-neural-networks/
A Beginner's Guide to the Mathematics of Neural Networks by ACC Coolen
http://citeseerx.ist.psu.edu/viewdoc/download?
doi=10.1.1.161.3556&rep=rep1&type=pdf
Explained: Neural networks [MIT]
http://news.mit.edu/2017/explained-neural-networks-deep-learning-0414
ML TUTORIALS
Tensorflow for Poets
https://codelabs.developers.google.com/codelabs/tensorflow-for-poets/
Building powerful image classification models using very little data
https://blog.keras.io/building-powerful-image-classification-models-using-
very-little-data.html
Learning AI if You Suck at Math
https://hackernoon.com/learning-ai-if-you-suck-at-math-8bdfb4b79037](https://image.slidesharecdn.com/introtocoreml-170627060159/85/Machine-Learning-on-iOS-with-Core-ML-29-320.jpg)
Machine Learning on iOS with Core ML
- 1. 🤖 Fun Machine Learning 🤖 🍎 Now on iOS 🍎 🎉 Yay 🎉
- 2. Machine Learning on iOS with Core ML @josephhosford Developer w/ Music Loft Apps musicloftapps.com josephhosford.com
- 3. I make music apps And dabble in image classification in ML.
- 4. 🤖 Machine Learning 🤖 🍎 ML on iOS 🍎 Two Parts
- 5. What is machine learning? 🤖 Great overview video https://www.youtube.com/watch?v=NOjUaY0Qn3g
- 6. 🤖 “Machine Learning is about making decision based on trial and error …a more application oriented version of statistics. The basic idea is to make decisions based on certain amount of information you have.“ -Anonymous via Quora
- 7. 🤖 Let us break down some ML terms: “ML got levels to it” Kendrick LaML Model Neural Networks Supervised & Unsupervised Learning Common Tools & Libraries
- 8. 🤖 Model
- 9. 🤖 Neural Networks If you want to learn more about the math behind NNs: read ACC Coolen’s paper (link in the resources) Modeled loosely on the human brain,a neural net consists of thousands or even millions of simple processing nodes that are densely interconnected. - MIT For a more detailed and basic overview: “A Quick Introduction to Neural Networks” (link in the resources)
- 10. 🤖 Supervised & Unsupervised Learning Supervised algorithms can apply what has been learned in the past to new data. Unsupervised algorithms can draw inferences from datasets.
- 11. 🤖 Common ML Tools & Libraries
- 12. 🤖 Let us walkthrough building a basic image classification ML model with Tensorflow HT: Google Codelabs “Tensorflow for Poets” and Siraj Raval on YouTube
- 13. 🤖 Conclusion: Machine Learning is totes powerful… a bit complicated… and often requires training and patience.
- 14. 🤖 Note: Machine Learning isn’t perfect.
- 15. Machine Learning on iOS with Core ML 🍎
- 16. What is Core ML? 🍎 A Framework that allows us to use ML Models in iOS 11 (an inference engine) Built on low-level tech (e.g.Metal & Accelerate) Utilizes ML Models with a new .MLModel file Ability to convert models from popular ML libraries
- 17. Supported Libraries by CoreML 🍎
- 18. The ML Stack on iOS 🍎🍎
- 21. 🍎🍎
- 22. Core ML The Good 🍎🍎 Very little iOS code Easy to implement Potential for powerful mobile experiences ML Tools and .MLModel Filetype are open source
- 23. 🍎🍎 Requires lot of ML training Only supports supervised learning Not a lot of easy / beginner resources to help in creating Core ML models Almost all ML libraries are in Python making it harder to just start experimenting Core ML The Bad
- 24. 🍎🍎 No on device training of the model .MLModel file is unencrypted Potential for huge file sizes No direct Tensorflow support (yet) Core ML The Ugly For more detailed criticism see: “Why Core ML will not work for your app (most likely)” (link in the resources)
- 25. 🍎🍎 Core ML Models aren’t the only way to do ML on iOS You can roll your own ML with Metal Performance Shaders and Accelerate. There are open sources libraries that implement ML on iOS. Check out Alex Sosnovshchenko’s frequently updated list: “Machine Learning for iOS” (link in resources)
- 26. Thanks for your time! @josephhosford
- 27. RESOURCES
- 28. IOS ML Machine Learning for iOS (Curated ML list of resources for iOS developers) http://alexsosn.github.io/ml/2015/11/05/iOS-ML.html CORE ML Core ML Product Page https://developer.apple.com/machine-learning/ Core ML Python Tools https://pypi.python.org/pypi/coremltools Why Core ML will not work for your app (most likely) http://alexsosn.github.io//ml/2017/06/09/Core-ML-will-not-Work-for-Your- App.html
- 29. NEURAL NETWORKS A Quick Introduction to Neural Networks https://ujjwalkarn.me/2016/08/09/quick-intro-neural-networks/ A Beginner's Guide to the Mathematics of Neural Networks by ACC Coolen http://citeseerx.ist.psu.edu/viewdoc/download? doi=10.1.1.161.3556&rep=rep1&type=pdf Explained: Neural networks [MIT] http://news.mit.edu/2017/explained-neural-networks-deep-learning-0414 ML TUTORIALS Tensorflow for Poets https://codelabs.developers.google.com/codelabs/tensorflow-for-poets/ Building powerful image classification models using very little data https://blog.keras.io/building-powerful-image-classification-models-using- very-little-data.html Learning AI if You Suck at Math https://hackernoon.com/learning-ai-if-you-suck-at-math-8bdfb4b79037
Editor's Notes
- #13: curl -O https://raw.githubusercontent.com/tensorflow/tensorflow/r1.1/tensorflow/examples/image_retraining/retrain.py tensorboard --logdir training_summaries & python retrain.py \ --bottleneck_dir=bottlenecks \ --how_many_training_steps=500 \ --model_dir=inception \ --summaries_dir=training_summaries/basic \ --output_graph=retrained_graph.pb \ --output_labels=retrained_labels.txt \ —image_dir=star_wars