Google Cloud - Google's vision on AI
- 1. Google's vision on AI Big Data Expo 19th September 2018 Rokesh Jankie [email protected]
- 3. 4th of October 2016
- 4. 4th of October 2016
- 5. Putting things in perspective
- 6. Machine learning is learning from examples and experience
- 7. Let’s try some human-powered image detection
- 8. How would we do this without ML?
- 9. How would we do this without ML?
- 10. How would we do this without ML?
- 11. What about a dog... and a mop? Easy, right?
- 12. Not so fast... CC-BY-SA-2.5 Wikimedia Commons https://commons.wikimedia.org/wiki/File:Komondor_Westminster_Dog_Show_crop.jpg CC-BY-2.0 Wikimedia Commons https://commons.wikimedia.org/wiki/File:2014_Westminster_Kennel_Club_Dog_Show_(12487315865).jpg CC-BY-2.0 Petful https://www.flickr.com/photos/petsadviser-pix/16395099127 CC-BY-SA-2.0 Jeffrey Beall https://www.flickr.com/photos/denverjeffrey/6903790333
- 14. How Can You Get Started with Machine Learning? • Three ways, with varying complexity: • Use a Cloud-based or Mobile API (Vision, Natural Language, etc.) • Use an existing model architecture, and retrain it or fine tune on your dataset • Develop your own machine learning models for new problems Moreflexible,butmoreeffortrequired
- 16. In case you were wondering... CC-BY-SA-2.5 Wikimedia Commons https://commons.wikimedia.org/wiki/File:Komondor_Westminster_Dog_Show_crop.jpg CC-BY-2.0 Wikimedia Commons https://commons.wikimedia.org/wiki/File:2014_Westminster_Kennel_Club_Dog_Show_(12487315865).jpg CC-BY-2.0 Petful https://www.flickr.com/photos/petsadviser-pix/16395099127 CC-BY-SA-2.0 Jeffrey Beall https://www.flickr.com/photos/denverjeffrey/6903790333
- 17. What is TensorFlow • Open source Machine Learning library • Especially useful for Deep Learning • For research and production • Apache 2.0 license
- 19. Jeff Dean!
- 20. History of Open Source @Google
- 21. You mentioned Deep Learning? What ?! • The first question to answer: What’s a Neural Network ? • Inspired by Biology: • Two flavors: Supervised an Unsupervised
- 22. This is what we are trying to solve f(x) = y
- 23. Neural Networks • Many kinds op NNs
- 24. What happened in the last decade… • Algorithms: This area has seen some improvements, but most of the early wins came from fairly old ideas. Now that Deep Learning is showing success we are seeing some good advances as well. • Datasets: Training large networks is hard without large enough datasets. MNIST can only go so far in pushing the limits. Having datasets like ImageNet has really helped pushed the state of the art in vision. • Compute: the biggest game changer in recent years.
- 25. 2012
- 28. TensorFlow History: The Cat This was in 2012 It took 16,000 computers to identify a cat!
- 29. Linear Regression & Classification
- 33. what just happened in Linear Regression ?
- 34. We need more than just software
- 36. Confidential & ProprietaryGoogle Cloud Platform 36 11.5 petaflops per pod
- 41. On device machine learning
- 42. Examples • Image2Text • “Lip reading” • Human-like sound: WaveNet • Skin cancer detection • AI Experiments by Google • Learn how to fly
- 44. Google Duplex
- 46. Google Lens
- 47. Possibilities…many, but I want to highlight one specifically
- 49. Solution space
- 50. – Aja Huang “The possibilities and power are innumerable”
- 54. What’s Next ?
- 56. Google AI Research Centers
- 57. Google AI Research Centers
- 60. Google & IoT Edge TPU
- 63. 4th of October 2016
- 64. Final Note, by the man himself!