Intro to Machine Learning for Android developers

Transcript

1. Intro to ML for Android developers November 2016
2. 2010s : emergence of Machine Learning 1999 : neural networks used for handwritten digits recognition
3. 2010s : emergence of Machine Learning 1999 : neural networks used for handwritten digits recognition 2012 : @Google, 16k processors, 10M Youtube videos, 1 week
4. 2010s : emergence of Machine Learning 1999 : neural networks used for handwritten digits recognition 2012 : @Google, 16k processors, 10M Youtube videos, 1 week Able to find cats
5. 2010s : emergence of Machine Learning • Self-driving cars • Human interaction : ◦ Handwriting ◦ Speech ◦ Natural language • OCR • Image recognition • Information retrieval • Artificial personal assistants • Recommendations systems • Drones • Game playing • ...
6. Machine Learning
7. Explaining Machine Learning Machine learning is the idea that there are generic algorithms that can tell you something interesting about a set of data without you having to write any custom code specific to the problem. Instead of writing code, you feed data to the generic algorithm and it builds its own logic based on the data.
8. Explaining Machine Learning Different categories : • Supervised learning ◦ continous answer : regression ex : estimate a price, a volume, ... ◦ discrete answer : classification ex : spam detection, cancerous tumors • Unsupervised learning • Reinforced learning continuous environment and constant feedback, learns by itself
9. Unsupervised learning
10. Linear regression
11. Intuitions from linear regression • algorithm is generic, results depends on data • system is both the algorithm and the data • starts with a hypothesis about how we can represent the data (for linear regression : a straight line) • only as good as your data • can deal poorly with outliers • lots of calculation to learn, but very fast to apply (can run on mobile)
12. Tensorflow Playground
13. Deep learning
14. State of the art : Google
16. SyntaxNet and Parsey McParseface
17. TensorFlow ML framework Opensourced in 2016 New standard
18. Implications for mobile
20. Interactions between mobile and ML • Mobile as a source of data
21. Assistant API d.android.com/training/articles/assistant.html
22. Interactions between mobile and ML • Mobile as a source of data • Mobile as UI ◦ still need UI to access services (even AI services) ◦ bots are not for everyone : alternative UI for “power users” • ML to enrich mobile apps (offline ML)
23. ML offline
24. Applications of NLP at Quora - automatic grammar correction - question quality - duplicate question detection - related question suggestion - topic biography quality (= qualifications of writer) - topic labeler (from “science” to narrow topics like “tennis Courts in Mountain View”) - search - answer summaries - automatic answers wiki - hate speech/harassment detection - spam detection - question edit quality
25. Questions? November 2016
27. Overfitting Need to split the data into : - training set (60%) - cross-validation set (20%) - evaluation set (20%)
28. SyntaxNet and Parsey McParseface Parsey McParseface can correctly read: • The old man the boat. • While the man hunted the deer ran into the woods. • While Anna dressed the baby played in the crib. • Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo. It makes mistakes on: • I convinced her children are noisy. • The coach smiled at the player tossed the frisbee. • The cotton clothes are made up of grows in Mississippi. • James while John had had had had had had had had had had had a better effect on the teacher