they re good dogs
play

They're Good Dogs A gentle introduction to Core ML and Vision - PowerPoint PPT Presentation

They're Good Dogs A gentle introduction to Core ML and Vision Andrew Harvey (@mootpointer) A Question For You What we will not cover In depth machine learning Cats Bleeding edge real - world implementations What we will cover


  1. They're Good Dogs A gentle introduction to Core ML and Vision Andrew Harvey (@mootpointer)

  2. A Question For You

  3. What we will not cover • In depth machine learning • Cats • Bleeding edge real - world implementations

  4. What we will cover • Dogs • The basics of the tools available • More dogs • What you could do with this stu ff

  5. Pro Tip

  6. "The best part of WWDC is the labs."

  7. "The best part of WWDC is the labs ! "

  8. Enough dogs. Show me the code (Said no one ever.)

  9. xkcd.com/1425

  10. Let's break it down

  11. Your App

  12. Domain Specific

  13. Vision

  14. Vision With a pixel buffer VNImageRequestHandler(cvPixelBuffer: imageBuffer, options: [:])

  15. Vision Or a CGImage VNImageRequestHandler(cgImage: image, options: [:])

  16. Natural language processing

  17. GameplayKit

  18. Core ML

  19. Core ML The fun stuff! model.prediction(from: featureProvider)

  20. MLFeatureProvider

  21. Metal Performance Shaders

  22. Accelerate and BNNs

  23. If you're starting out You probably want Core ML

  24. Any Questions?

  25. That's enough code. More dogs!

  26. Let's talk about Adversarial Examples

  27. Here's one I prepared earlier

  28. (Live demo ensues)

  29. Ecosystem

  30. You still have to train your own model

  31. Keras keras.io

  32. Caffe caffe.berkeleyvision.org

  33. scikit-learn scikit-learn.org

  34. coremltools developer.apple.com/machine-learning

  35. So what?

  36. Privacy

  37. Latency

  38. Offline

  39. "If you can avoid a network round trip, you probably should."

  40. Core ML + ARKit = !

  41. Where to from here?

  42. Learn you some acronyms CNNs, LSTMs, RNNs. (Oh my!)

  43. xkcd.com/1838

  44. fast.ai

  45. Make your own. developer.apple.com/machine-learning

  46. Understand your features

  47. Do No Evil

  48. Make fun things!

  49. Thanks! (I'm hiring!) andrew@canceraid.com

Recommend


More recommend