Machine Learning - Intermediate Track

As google's ExploreML facilitator conducted Intermediate track of ML at Women Techmaker, Baroda. This workshop conducted at the Mastercard office at Baroda.

Transcript

1. Explore Machine Learning Krupa Galiya Explore ML Facilitator
2. https://www.autodraw.com/ OR https://magic-sketchpad.glitch.me AutoDraw is a new kind of drawing tool that pairs the magic of machine learning with drawings from talented artists to help everyone create anything visual, fast. 
3. Move Mirror https://experiments.withgoogle.com/collecti on/ai/move-mirror/view/mirror
5. if object.height > 10: do x if object.color is blue: do y if object.numberOfLegs > 2: do z ... Frame the goal of the product Refine until goals are met Designers and engineers develop flow and logic Quickdraw: Rule-based
6. Quickdraw: Machine Learning Frame the goal of the product Refine until goals are met Train a model using examples
8. Machine Learning is a type of Artificial Intelligence that provides computers with the ability to learn without being explicitly programmed. Learn from Data. it is the study of computer algorithms that improve automatically through experience.
9. Rules-based approach Machine Learning Alphabetising a list of song titles Rules-based Which Approach to Use?
10. Rules-based approach Machine Learning Ranking Web Search Results Rules-based Machine Learning Which Approach to Use?
11. Rules-based Machine Learning Predicting Housing Prices based on Location Machine Learning Which Approach to Use?
12. Processing online payments Rules-based approach Machine Learning Rules-based Which Approach to Use?
13. Classifying an object in a photo Rules-based Machine Learning Machine Learning Which Approach to Use?
14. Each has its benefits Rule-based Approach Machine Learning – Rules are defined – Improvements come from algorithms and network – Learns patterns from data – Improvements can from additional data Recap
15. Idea to Implementation
16. Collect Data Train and Test Model Define Objective Predict and Evaluate Focus on User Machine Learning Process
17. Artificial Intelligence Machine Learning Deep Learning
18. MAMMAL WILDLIFE LION Classification
19. Regression
20. Clustering
21. Classification Clustering Regression Sequence Prediction Style Transfer Classification Labeling email as spam or not-spam Type of Machine Learning?
22. Classification Clustering Regression Sequence Prediction Style Transfer Type of Machine Learning? Determining workout activity based on phone movement. Classification
23. Classification Clustering Regression Sequence Prediction Style Transfer Type of Machine Learning? Predicting the strength of a password Regression Classification
24. Classification Clustering Regression Sequence Prediction Style Transfer Type of Machine Learning? Identifying famous landmarks in a photo Classification
25. Classification Clustering Regression Sequence Prediction Style Transfer Type of Machine Learning? Predicting the quality score for an advertisement Regression
26. Classification Clustering Regression Sequence Prediction Style Transfer Type of Machine Learning? Estimating arrival time based on time of day and traffic. Regression
27. Neural Network
28. Perceptron • Perceptron is a single layer neural network and a multi-layer perceptron is called Neural Networks. • Perceptron is a linear classifier (binary). Also, it is used in supervised learning.
29. Let’s try to understand by Visualization https://playground.tensorflow.org/
30. What is TensorFlow?
31. TensorFlow is a free and open-source software library for dataflow and differentiable programming It is a symbolic math library, and can be used for machine learning applications such as neural networks.
32. Why TensorFlow?
33. TensorFlow 2.0 Architecture
35. Demos
36. Text Classification That was the best movie ever! Positive 10 Everyone but me liked the movie Negative 5 That was the worst movie ever! Negative 10 I would rate it 7 out of 10 Positive 7 It was ok ? 2 Positive Negative
37. 1. bit.ly/2WZPsEn 2. Open the exercise in Colab 3. Sign in with your Google account 4. Activate the Runtime 5. Run code Text Classification Coding Activity
38. Transfer Learning Updated Explore Data Retrain Classification Layers Predict Evaluate Reuse Pretrained Model
40. What is a pre-trained model?
41. A pre-trained model is a saved network that was previously trained on a large dataset, typically for a large-scale image-classification task.
42. Transfer Learning Source Model Source Data Output Labels Source Model Output Labels New Data (different domain)
43. 1. bit.ly/32yjMXM 2. Open the exercise in Colab 3. Sign in with your Google account 4. Activate the Runtime 5. Run code Image Classification Coding Activity
44. Resources to start ML
45. https://goo.gl/3L9cGu
46. 1. bit.ly/AttendenceExploreML 2. bit.ly/exploreml-s-post 3. bit.ly/exploreml-s-pre Fill this forms