Ce diaporama a bien été signalé.
Nous utilisons votre profil LinkedIn et vos données d’activité pour vous proposer des publicités personnalisées et pertinentes. Vous pouvez changer vos préférences de publicités à tout moment.

Machine Learning Introduction

176 vues

Publié le

An introduction to machine learning. From my bootcamp S01E01.

  • Soyez le premier à commenter

Machine Learning Introduction

  1. 1. Why is machine learning exciting ? Self driving cars Voice recognition Alpahgo
  2. 2. Speaker : Benjamin Ejzenberg @TheBenimou
  3. 3. @TheBenimou Benjamin Ejzenberg https://thebenimou.github.io/
  4. 4. @TheBenimou Benjamin Ejzenberg https://thebenimou.github.io/ Senior @EY Data enthusiast Digital addict Meme maker CDO Hacker Jazz improviser
  5. 5. How would we do ML without ML ?
  6. 6. How would we do ML without ML ?
  7. 7. Machine Learning APIs by Example (Google I/O '17)
  8. 8. Machine Learning APIs by Example (Google I/O '17)
  9. 9. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules
  10. 10. Study the problem Train ML algorithm Analyze errors Write rules
  11. 11. Study the problem Train ML algorithm Analyze errors Write rules
  12. 12. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules
  13. 13. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules • IF shape is round, DO this • IF color is orange, DO this • IF texture is this, DO this • … • (very long & complex list)
  14. 14. dy the oblem Train ML algorithm Evaluate Launch Analyze errors Write rules
  15. 15. Train ML algorithm Evaluate Launch Analyze errors Write rules Test the model with n images
  16. 16. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules Then, two options …
  17. 17. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules The model does not perform well, let’s iterate again.
  18. 18. Train ML algorithm Evaluate Launch Analyze errors Write rules The model is good ? Let’s deploy it to production !
  19. 19. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules
  20. 20. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules Since the problem is not trivial, your program will likely become a long list of complex rules—pretty hard to maintain.
  21. 21. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules Since the problem is not trivial, your program will likely become a long list of complex rules—pretty hard to maintain.
  22. 22. Machine Learning APIs by Example (Google I/O '17)
  23. 23. Machine Learning APIs by Example (Google I/O '17)
  24. 24. Machine Learning approach
  25. 25. Machine Learning is the science (and art) of programming computers so they can learn from data.
  26. 26. ML […] gives computers the ability to learn without being explicitly programmed. Arthur Samuel, 1959
  27. 27. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules
  28. 28. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules
  29. 29. Study the problem Train ML algorithm Evaluate Launch Analyze errors Write rules This time, we will note write rules ourselves
  30. 30. Study the problem Train ML algorithm Evaluate Launch Analyze errors Data
  31. 31. Study the problem Train ML algorithm Evaluate Launch Analyze errors Data
  32. 32. Study the problem Train ML algorithm Evaluate Launch Analyze errors Data
  33. 33. Why learn Python in 2017 ?
  34. 34. Hello world in Python vs. C & Java Easy to learn Clean syntax Comprehensive standard library Excellent documentation Immediacy of writing and running a script
  35. 35. Let’s practice now.

×