????????? ?????????????(????) ??#1-??
?
25 likes?4,128 views
??? ??? ????? ?? --> https://pg.mycampus.io/ #CNN #??? #keras #tensorflow #kaggle
????????? ?????????????(????) ??#1-??
- 1. ???? J ????? ??? ?? ????? @ ??????, ?????? ??? ?? ¨C ???/???? ?? ?? ??? @ ?????? ??? ??/??? daehee@slowcampus.com http://medium.com/@slowcampus http://slowcampus.com Just Start up your Deep Learning for your Future !
- 2. ???? J ????? ??? ?? ????? @ ??????, ?????? SLOW? SW(?????) ??? 2 ???? ??? ??? ????? ???? http://slowcampus.com ??? http://medium.com/@slowcampus
- 3. ???? J ????? ??? ?? ????? @ ??????, ??? ??: ??? Tensorflow & Keras 1? ?? ? ??? ??? ?? ? ??? ?? ?? ? ??? ?? ?? 3 ??? ??? ???
- 4. ???? J ????? ??? ?? ????? @ ??????, ????? ??? ???, ?? ???? ??Goal ?? !! 4 ??? ??/?? ?? ???? ????? ???? ???? ??? ?? ???? ???? ?? ?????? ???? ?? (???) ??? DB, ???? Goal: ???? ?? ???
- 5. ???? J ????? ??? ?? ????? @ ??????, ???? ?? ? ?1? ?????? ¨¤ ???? ??? ?? ?? ?? (1980?~2010?) ? ?2? ?????? ¨¤ ?????/??? ?? ????? ????? ?? (2010?~2040?) ??? ??(HW)? ????, ????? ??? ?????, ??? SW??? ??/???? ?? ?. ? ?? ???? ?? ??? ????. ?????? ??, ?????? ??? ?????? ??? ????? ?? ???. ?? ??? ?? ???? ?? ?? ???? ??? ??? ??? ??. 5 Data Information Knowledge
- 6. ???? J ????? ??? ?? ????? @ ??????, ???/???? ?? ?? ???? ? ??? ??? ???? ? ?? ?? ML Researcher ? ??? ???? ???? ???? ?? ??? ?? ML Engineer ? ?????/??? ??? ??/?? Data Engineer 6
- 7. ???? J ????? ??? ?? ????? @ ??????, ????? ?? ÅДà Judgement, Decision 7 1 .??? ???? ??? ?? ?? ?? ??? ??. 2 .?? ??? ??? ?? ???? ???? ??? ?? ??.
- 8. ???? J ????? ??? ?? ????? @ ??????, ??? ???? ????? ??? ??? ???? ??/??? ????? ? ?? (Rule) ?? ? ??(Example) ?? ¨C ?? ? ??? ?? 8 ¨¤ ??? Data Driven Decision ??? ?? ?? ?
- 9. ???? J ????? ??? ?? ????? @ ??????, ??? ?? ??(??? ?? ?)? ?? 9 ??(??) ?? ?? ?? ?? ?? ??? ????? ? ? ??? ???? ?? ? ??? ? ???? Logistic Yes/No, True/False Logistic regression ???? ???? ?? ?? ???? (Discrete) Classification ?? ???? ??, ??? ??, ????? ?? ? ?? ??? (Continuous) Regression, Prediction, ¡
- 10. ???? J ????? ??? ?? ????? @ ??????, ?? ?? - regression 10 ??? ?(point, ???)???? Input X? output Y ??? ???(??)? ?? ?
- 11. ???? J ????? ??? ?? ????? @ ??????, ???? ¨C Classification (??) Input X? ?? label(output Y, discrete values) ? ??? ? CAT (1) DOG (2) ?? X ?? Y
- 12. ???? J ????? ??? ?? ????? @ ??????, ???? - Classification Input X? ?? label(output Y) ? ??? ? 0 1 2 ... 9 ?? X ?? Y
- 14. ???? J ????? ??? ?? ????? @ ??????, Fit & Modeling ???? ??? ?? 5~6 ??? ??? ??(???)?? ?? 14
- 15. ???? J ????? ??? ?? ????? @ ??????, Fit & Modeling ???? ?? ?? 1?? ?? ??? ???? ?? 15
- 16. ???? J ????? ??? ?? ????? @ ??????, Fit ?? ?? ?? ???? ?? ?? ?? ???? 16
- 17. ???? J ????? ??? ?? ????? @ ??????, ? ? ?? ?? ?? ???? ?? ? ? ???? ?? ??? ???? ?? 17 Fitting
- 18. ???? J ????? ??? ?? ????? @ ??????, ? ??? ??? ??? 18
- 19. ???? J ????? ??? ?? ????? @ ??????, Exponential fit: y = aebx ? ???(?)? ?? fit ?? a, b ?? ??? ? ????? 2?? ? ??. (a, b) 19
- 20. ???? J ????? ??? ?? ????? @ ??????, Power fit: y = axb 20 ? ???(?)? ?? fit ?? a, b ?? ??? ? ????? 2?? ? ??. (a, b)
- 21. ?? ?? ? ???? ?? (????? ???? ?? ???) 21
- 22. ???? J ????? ??? ?? ????? @ ??????, ¨C H.B. BARLOW ¡°Intelligence is the art of good guesswork¡±
- 23. ???? J ????? ??? ?? ????? @ ??????, ¨C ? ? ? J ¡°???? ?? ???? ???? ??, ??? ???? ???? ?? ?¡±
- 24. ???? J ????? ??? ?? ????? @ ??????, ??? ¨C fitting ? ?? ?? ?? ?? 24 ??? ??? ? ??? ?????, ??? ??? ??? ?? ?? ??? ????.
- 25. ???? J ????? ??? ?? ????? @ ??????, ??? ? Input ¨C output ?? ??/??/??/??? ?? ? ???? ??? ???? 25 * ???? ???? ????? ??? ??? ???? ??? ??? ??. X? ???? feature? ??? ? ?? ??? ????. ????: ?? x? ?? ? ??? y? ???? ??.
- 26. ???? J ????? ??? ?? ????? @ ??????, ????? ??? ?? ? (??) 26 ?? ? ???(X) ???(Y) Features Class, Label Pattern Function Vectors Relation ??/??? ??? (training/test) Probability Function
- 27. ???? J ????? ??? ?? ????? @ ??????, ??? ?Input ¨C output ???? 27 ? ?
- 28. ???? J ????? ??? ?? ????? @ ??????, ??? ¨C ?? ? ?? ? Input ¨C output ??? ?? 28 0.0 0.0 0.0 ¡. 0.0 0.0 0.0 ¡. 0.0 0.0 0.0 ¡. 0.9 0.8 0.1 ¡ 0.2 0.3 0.5 ¡ 0.2 0.3 0.5 ¡ ¡ ?? ??
- 29. ???? J ????? ??? ?? ????? @ ??????, ????? (2?1?) ? ???(??)? 2??? ?? 2? ???? ? ???? ??? ??? ????(??)? ?? ?? 29
- 30. ???? J ????? ??? ?? ????? @ ??????, ????? (3?1?) 30 ? X (x, y, z): ??? ? A, B? ?? ???? ? A? ??? Weight ? B? ??? Bias
- 31. ???? J ????? ??? ?? ????? @ ??????, 31 ????? (N?1?) ? X (x1, x2, ¡, xn): ??? ? A, B? ?? ???? ? A? ??? Weight ? B? ??? Bias
- 32. ???? J ????? ??? ?? ????? @ ??????, ??? ¨C ?? ?? ?? ? ?? output? ??? output? ??(Erorr, Loss, Cost)? ? ????? ????? ? (minimize) ? Iteration == Loop == Epoch 32 0.9 0.8 0.1 ¡ 0.2 0.3 0.5 ¡ 0.2 0.3 0.5 ¡ ¡ X Y¡¯Y ?? ???? E = Y ¨C Y¡¯ ????(W, b) ??
- 33. ???? J ????? ??? ?? ????? @ ??????, ??? ?? ? ??(Loss, Cost)? ????? ?? ?? ?? ???? (weights)? ????? ??? ??? ??? ???. 33 X: ?? ?? ??? Y: ?? ???(???) Y¡±: ??? ?? ??? ??(MSE, cost, loss)?? ????
- 34. ???? J ????? ??? ?? ????? @ ??????, Training, Validation, Test ? Training ¨C ??? ? Validation ¨C ?? ????: ? ? ??? ???? ?? ? Testing ¨C ?? ????, ?? ??? ??? ?? ?? 34 ??? ?? ??? ??? (Training) ???? (Testing) ?? ??? (Validation) 80%, 10%, 10%
- 35. ???? J ????? ??? ?? ????? @ ??????, ??(training)? ??? 35 training(X, Y) pair: Data for Training Y¡¯ - ??? ??? ????? ????? (weights, biases) ?? (???? ??) Save X¡¯: Data for Test test ????? ????? (weights, biases) Load Threshold & Decision
- 36. ?? ?? ?? 36
- 37. ???? J ????? ??? ?? ????? @ ??????, ??? ??? ?? 37 ? ??????? ???? ??, ??? ??? ??? ??
- 38. ???? J ????? ??? ?? ????? @ ??????, ?? ??? ?? 38 Supervised Unsupervised Reinforcement Learning ???? ?? (X, Y)? ?? ???? ?? Guess & Measure ??? ??? ? ? ??? ?? ? ?: ??, ??
- 39. ???? J ????? ??? ?? ????? @ ??????, Supervised ?? ??) Lunit 39
- 40. ???? J ????? ??? ?? ????? @ ??????, ??) ?? 40
- 41. ???? J ????? ??? ?? ????? @ ??????, Unsupervised??) Dable?? ??, ?? ?? 41
- 42. ???? J ????? ??? ?? ????? @ ??????, Unsupervised ??) Watcha ?? ??, ?? ?? 42
- 43. ???? vs ????? 43 http://www.kdnuggets.com/2016/05/implement-machine-learning-algorithms-scratch.html ? ???? ¨C ? ????? ¨C ??? ??? ??? ??
- 44. ??, ?? ???? ???? 44
- 45. ???? J ????? ??? ?? ????? @ ??????, ?????? ??? ????? ?? ????? ???. ????? ??? ???????? ????? ????? ??/????? ??? ???? ?? ? ?? ???? ? 45 ???? ???? ??? ??(Method) ??(Goal)
- 46. ???? J ????? ??? ?? ????? @ ??????, ???? == ??? ?? ??????? ??(???)? ????? ??/??/??/? ?/??? ???? ?. ¨¤ Learning from Data 46 Inductive Learning (learning by examples) automatic discovery of regularities in data through the use of computer algorithms and generalizing those into new but similar data
- 47. ???? J ????? ??? ?? ????? @ ??????, ?????? ? ?? ? ???? ?? ????? ??? ?? ??? ??? ?????? ???? ?? ? ??? ¨C ?, ?? ?? (Collaboration filtering) ? ??? ¨C ???? ?? (Associative Rules) ? ¡ 47
- 48. ???? J ????? ??? ?? ????? @ ??????, ???? ? ????? ??? ? ?? ? (?? ??/??/?? ????) ? ??, ??? ?? ? ??, ?? ???? ? ??? ? ?? ? ???? 48
- 49. ???? J ????? ??? ?? ????? @ ??????, ????? ????? ??? ??? ???? ?? ??? ??/?? ? ???? (??,??,??,???,???) ? ???? ??? (???) ? ?? (????, ????) ? ?? (??) ? IoT ¨C ??? ??? ?? ? ????+????? ? ??? ??,???
- 50. ???? J ????? ??? ?? ????? @ ??????, ?? ?? (Pattern Recognition) ? 1990?? ?? ? ????? ??? ???? ??? ???? ?? ?? ? ¡®??? ??¡¯ ??? ?? ??? ?? ? ¡®??? ??? ??/???? ???? ? ¡®?? ??¡¯? ???(??? ????)? ???? ? ? ?? 50
- 51. ???? ??? 51
- 52. ???? J ????? ??? ?? ????? @ ??????, ? ???/???? AI???? ??? ????? ??? ??????? 52 ??? ¨C ???? ??? ??? ???? ????? ! ??/??/??/???? ?? ??? ?? ??? ?? !! ???
- 53. ???? J ????? ??? ?? ????? @ ??????, ??? ?? ??? ???? ???? ??? ! 53 ???? ??? ???? 1.??? 2.???3.?? ??? ??,??,???? Data Scientist Data Engineer ??? ??, ??
- 54. ????? ??? vs ?? ?? ??? 54
- 55. ???? J ????? ??? ?? ????? @ ??????, 55
- 56. ???? J ????? ??? ?? ????? @ ??????, Statistics vs Machine Learning 56https://www.datasciencecentral.com/profiles/blogs/machine-learning-vs-traditional-statistics-different-philosophi-1 https://svds.com/machine-learning-vs-statistics/ ¡°Machine learning is for Computer Science majors who couldn¡¯t pass a Statistics course.¡±
- 57. ???? J ????? ??? ?? ????? @ ??????, Statistics vs Machine Learning 57https://www.datasciencecentral.com/profiles/blogs/machine-learning-vs-traditional-statistics-different-philosophi-1 https://svds.com/machine-learning-vs-statistics/ ¡°Machine learning is essentially a form of applied statistics¡±
- 58. ???? J ????? ??? ?? ????? @ ??????, Statistics vs Machine Learning 58https://www.datasciencecentral.com/profiles/blogs/machine-learning-vs-traditional-statistics-different-philosophi-1 https://svds.com/machine-learning-vs-statistics/ ¡°Machine learning is Statistics minus any checking of models and assumptions.¡±
- 59. ???? vs ??/?? ?? ? Deterministic ? ??, ?? ?? ?? ? ???? ????, ??? ? ? ?? ?? ?? ?? ?? ?? ? Stochastic (probabilistic + time) ? ????? ?? ?? ??? ?? ??? ??? ? ???? ?? ?? ?? ?? 59
- 60. ???? J ????? ??? ?? ????? @ ??????, Machine Learning (ML) Traditional statistics (TS) Goal: ¡°learning¡± from data of all sorts Goal: Analyzing and summarizing data No rigid pre-assumptions about the problem a nd data distributions in general Tight assumptions about the problem and data distri butions More liberal in the techniques and approaches Conservative in techniques and approaches Generalization is pursued empirically through t raining, validation and test datasets Generalization is pursued using statistical tests on the training dataset Not shy of using heuristics in approaches in se arch of a ¡°good solution¡± Using tight initial assumptions about data and the pr oblem, typically in search of an optimal solution unde r those assumptions Redundancy in features (variables) is okay, and often helpful. Preferable to use algorithms desi gned to handle large number of features Often requires independent features. Preferable to us e less number of input features Does not promote data reduction prior to learn ing. Promotes a culture of abundance: ¡°the mo re data, the better¡± Promotes data reduction as much as possible before modeling (sampling, less inputs, ¡) Has faced with solving more complex problems in learning, reasoning, perception, knowledge presentation, ¡ Mainly focused on traditional data analysis 60
- 61. ?? ?? vs ???? 61
- 62. ???? vs ???? ???? ? ???? ???? ????? ? ? ????? ?? ? ????? ??? ??? ??? ? ?? ???? ?. ?? ???. ???? ? ????? step-by-step ?? ?? ? ?? ??/??/??? ?? ????? ??? ? ????? ????? ?? 62
- 63. ???? J ????? ??? ?? ????? @ ??????, ???? ?? - sorting 63
- 64. ???? J ????? ??? ?? ????? @ ??????, ???? ?? - sorting 64
- 65. ???? J ????? ??? ?? ????? @ ??????, ???? ?? - sorting 65