Europython - Machine Learning for dummies with Python
2 likes1,021 views
The document provides an overview of machine learning, focusing on its importance, applications, and basic concepts with Python examples. It discusses the process of building machine learning models, including steps such as gathering data, selecting algorithms, training models, and evaluating predictions. Additionally, it highlights resources for learning machine learning, including MOOCs and libraries like Scikit-learn and TensorFlow.
Europython - Machine Learning for dummies with Python
- 1. Machine Learning for dummies with Python EUROPYTHON Javier Arias @javier_arilos
- 2. One day in your life July 2016
- 3. One day in your life Time to leave the office
- 4. One day in your life Tesla autopilot
- 5. One day in your life Playing music
- 6. One day in your life Your photos organized
- 7. One day in your life Machine Learning is here, it is everywhere and it is going to stay
- 8. About this presentation Why Machine Learning (ML) matters A journey on Machine Learning Some ML technologies and resources Some basic ML concepts, with code samples
- 9. Machine Learning is the next big thing
- 10. Are machines already intelligent?
- 11. Image-net challenge 2015: machines outperform people
- 12. Chess 1997: Deepblue defeats Kasparov
- 13. Game of Go 2016: AlphaGo wins world champion Lee Sedol
- 15. The journey
- 16. Learning about ML MOOC - Massive Open Online Courses Contents by the best universities and companies Udacity, Coursera, EdX
- 17. Udacity - Intro to Machine Learning Pattern Recognition for Fun and Profit - Very well organized contents - Python + sklearn - Free - At your own pace
- 18. Udacity - Intro to Machine Learning Pattern Recognition for Fun and Profit
- 19. Udacity - Intro to Machine Learning Pattern Recognition for Fun and Profit
- 20. What is Machine Learning?
- 21. Solving a complex problem somethingfeatures (data) prediction
- 22. First approach, programming tell the computer what to do in very tiny steps
- 23. programming does not scale for very complex problems... First approach, programming
- 24. Machine Learning show the computer some real world data the algorithm will learn from it
- 25. Machine Learning, implications we can train computers to do things we do not know how to do
- 27. ML example: character recognition Not-MNIST dataset Thousands of 28x28 grayscale images with labels
- 28. features x 1000s labels x 1000s F G F J ML step 1: get samples (training data)
- 29. ML step 2: choose an algorithm Linear regression Support Vector Mach. k-Means Decision Trees Random Forests Neural networks Convolutional NN Naive Bayes
- 30. ML step 3: train your algorithm features x 1000s labels x 1000sML algorithm F G F J
- 31. ML, last step: getting predictions ML algorithmfeatures (data) prediction D
- 32. Tricky Question How good are our predictions?
- 33. The Tools
- 34. The Tools: Python ● Opensource ● Expressive ● Interpreted, dynamically typed ● Widely used many different problems ● Batteries included: Notebook, Libraries
- 35. The Tools: sklearn ● Opensource, Python ● Wonderful documentation ● Support to full ML lifecycle: ○ Feature engineering ○ Algorithms ○ Validation ○ Datasets
- 36. A summary of ML process ● Get features (with labels) ● Choose and configure an algorithm ● Train your algorithm ● Do predictions ● Validate your results
- 37. train your model tr_ds, _, tr_lbl, _ = train_test_split(dataset, labels, train_size=size, random_state=17) clf = LogisticRegression() clf.fit(tr_ds, tr_lbl) # fit with train dataset and train labels train_ds test_ds dataset
- 38. make predictions pred = clf.predict(test_dataset) How good are our predictions?
- 39. accuracy test_predicions = clf.predict(test_dataset) acc = accuracy_score(test_labels, test_predictions)
- 40. 89% accuracy
- 41. Improving prediction results Training data Algorithm + config
- 43. Udacity - Deep Learning Take machine Learning to the next level ML branch based on algorithms that use multiple processing layers ● By Google ● Python and Tensorflow ● No wine for the moment :-(
- 44. The Tools: TensorFlow ● Opensource, Python ● Deep Learning ● Data flow graphs. ○ Nodes: mathematical operations ○ Edges: Tensors, multidimensional arrays
- 45. Simplest Neural Network weight b i a s S O F T M A X X Y PREDweight b i a s LAYER 1 LAYER 2 * + relu * + softmax
- 46. Deep Learning as a chain of operations
- 47. Let’s recap
- 55. Thank you for your attention