Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

AIMeetup #3: Data is oxygen for ML

141 views

Published on

Dima Boyko podczas AIMeetup #3 w Krakowie organizowanego przez 2040.io opowiadał o tym jak ważne są dane w machine learning.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

AIMeetup #3: Data is oxygen for ML

  1. 1. Data is oxygen forML Cracow, 8th December 2016
  2. 2. Hello, I’m Dima Boyko Dima Boyko Rails developer, Python developer, Data Scientist Software engineer @inFakt dima.boyko@infakt.pl /dimaboyko
  3. 3. What computer can do better than human? What humancan do better than computer?
  4. 4. What computer can do better than human? + What humancan do better than computer?
  5. 5. ” Field of study that gives computers the ability tolearn without being explicitly programmed 1959, Arthur Samuel What computer can do better than human? + What humancan do better than computer? Machine learning
  6. 6. Machine learning is not the future
  7. 7. Machine learning ALVIN YouTube https://youtu.be/ilP4aPDTBPE?t=39 1992 !
  8. 8. Machine learning
  9. 9. Boost ML can boost existing products by improving quality and usability ofsome modules Unlock Using ML can unlock newproduct use-cases Machine learning
  10. 10. Machine learning Drawbacks?
  11. 11. Machine learning Drawbacks? WOW / WTF ratio
  12. 12. Data Algorithm Insight Machine learning Usage model
  13. 13. Data Machine learning Usage model
  14. 14. Data
  15. 15. Using of data
  16. 16. Using of data Red Roof Inn
  17. 17. 2 to 3% of flights werecanceled Using of data Red Roof Inn
  18. 18. 500 daily Using of data Red Roof Inn
  19. 19. 90 000 passengers Using of data Red Roof Inn
  20. 20. Weather data Using of data Red Roof Inn
  21. 21. Using of data Red Roof Inn 10% more revenue during season
  22. 22. Using of data Los Angeles Police Department
  23. 23. Historical DATA Analysis & Prediction Reaction Using of data Los Angeles Police Department
  24. 24. Using of data Los Angeles Police Department
  25. 25. Using of data Los Angeles Police Department
  26. 26. 33% 21% Less thefts Less victims Using of data Los Angeles Police Department
  27. 27. Using of data UPS Cargo Delivery
  28. 28. 16,9M Delivered cargos daily 195 Countries around the globe Using of data UPS Cargo Delivery
  29. 29. Orion • Mathematical model for operations research • Huge processing power in realtime Using of data UPS Cargo Delivery
  30. 30. 13 000 Tons of exhausts less Using of data UPS Cargo Delivery 6M Litres less fuel usage during the year + Faster deliveries
  31. 31. Using of data inFakt Automated Accounting
  32. 32. ~50 000 Invoices booked monthly by accountants Using of data inFakt Automated Accounting
  33. 33. AutoAccounting Brief product history Using of data inFakt Automated Accounting
  34. 34. • • • Data from last year Scikit-learn Infrastructure AutoAccounting Classification 15% invoices 95% correct
  35. 35. • • Data from last year Infrastructure AutoAccounting Classification 55% invoices 95% correct
  36. 36. AutoAccounting Classification Keep it simple
  37. 37. 3% Wrong Inconsistent 8/10 Human mistake AutoAccounting
  38. 38. ~70% Invoices booked automatically AutoAccounting
  39. 39. ~70% Invoices booked automatically 600 / month Hours saved for creative work AutoAccounting
  40. 40. Know your DATA!
  41. 41. Auto What’s next?
  42. 42. Auto What’s next? #worldwide #vendor_independent #simple
  43. 43. What’s next? Open Source ? /OpenAutoX
  44. 44. Thanks! Any questions? Dima Boyko Software engineer @ inFakt dima.boyko@infakt.pl /dimaboyko

×