![Machine Learning – How?
Raw data preprocessed to obtain a feature vector,
X, that adequately describes all the relevant features
for classifying examples
Each x is a list of (attribute, value) pairs. Example:
X = [Person:Susan, EyeColor:Brown, Age:40, Sex:Female]
The number of attributes is fixed
Each attribute has discrete or continuous values
An example can be interpreted as a point in an n-
dimensional feature space, where n is the number
of attributes
19](https://image.slidesharecdn.com/module61-220801164928-08d874c3/85/module-6-1-ppt-19-320.jpg)
module 6 (1).ppt
- 2. What is Machine Learning? Learning from Data Learns the relationship between the variables of a system (input, output and hidden) from direct samples of the system (data) Plenty of data and computational power can replace rule- based models to probabilistic data-driven modes Learning from Big Data (scalable Learning) Smart / Autonomous / Intelligent systems
- 3. 3
- 4. Why Machine Learning? There is no need to “learn” to calculate payroll etc…. No Machine Learning needed when the relationships between all system variables (input, output, and hidden) is completely understood! – known function This is NOT the case for almost any real world system!
- 5. A Generic Real World System System … … 1 x 2 x N x 1 y 2 y M y 1 2 , ,..., K h h h 1 2 , ,..., N x x x x 1 2 , ,..., K h h h h 1 2 , ,..., K y y y y Input Variables: Hidden Variables: Output Variables:
- 6. Machine Learning – Why? When human expertise does not exist Humans are unable to explain their expertise (speech recognition) Some tasks cannot be defined well, except by examples (e.g., recognizing people). Correlations and patterns can be hidden within large amounts of data. 6
- 7. Machine Learning – Why? Amount of knowledge available about certain tasks might be too large for explicit encoding by humans (e.g., medical diagnostic). Environment / solution change over time. Solution needs to be adapted to particular cases Set of all possible behaviors given all possible inputs is too large Process must generalize from the finite set of examples to produce an output / decision in new cases. 7
- 8. Learning from Data Data may belong to any domain such as text, image, video, speech, Bio informatics etc. Few Examples: Learning to recognize spoken words Learning to drive an autonomous vehicle Learning to classify new structures Learning to play games 8
- 9. Driverless Vehicle • Learning to drive an autonomous vehicle – Associate steering commands with image sequences Google Prototype Task T: driving on public, 4-lane highway using vision sensors Perform measure P: average distance traveled Training E: sequence of images and steering commands recorded while observing a human driver
- 10. Handwritten character recognition • It is very hard to say what makes a “2” • Wide variability of same numeral • Handcrafted rules will result in large no of rules and exceptions • Better to have a machine that learns from a large training set
- 11. Face Recognition Training examples of a person Test images
- 12. Machine Learning Tasks Recognition / Classification Clustering Prediction / Regression Anomaly detection Retrieval 12
- 13. Other Applications Recognizing patterns: Speech Recognition Facial identities or facial expressions Handwritten or spoken words Medical images Recognizing anomalies: Unusual sequences of credit card transactions Unusual patterns of sensor readings in a nuclear power plant Video surveillance Prediction: Future stock prices Sales prediction Information Retrieval
- 16. Machine Learning – How? Supervised learning Training data includes desired outputs classification, regression, outlier detection Unsupervised learning Training data does not include desired outputs Grouping similar instances - clusters; Big Data don’t have labels. Semi-supervised learning Training data includes a few desired outputs Reinforcement learning Rewards or penalty from sequence of actions After a set of trial-and error runs, learns best policy - sequence of actions that maximize the total reward.
- 17. Machine Learning - Where? Handwriting Recognition x: Data from pen motion. f(x): Letter of the alphabet. Disease diagnosis x: Properties of patient (symptoms, lab tests) f(x): Disease (or recommended therapy) Face recognition x: Bitmap picture of person's face f(x): Name of the person. Spam Detection x: Email message f(x): Spam or not spam. So many……… 17
- 18. Machine Learning – How? Data Representation Choice of Similarity or Distance measure Choice of Learning Algorithm 18
- 19. Machine Learning – How? Raw data preprocessed to obtain a feature vector, X, that adequately describes all the relevant features for classifying examples Each x is a list of (attribute, value) pairs. Example: X = [Person:Susan, EyeColor:Brown, Age:40, Sex:Female] The number of attributes is fixed Each attribute has discrete or continuous values An example can be interpreted as a point in an n- dimensional feature space, where n is the number of attributes 19
- 20. 20 Machine Learning – Techniques • Decision Tree induction Tree is constructed in a top-down recursive divide-and-conquer manner Simple rules Overfitting • Bayesian classifier Statistical classifier performs probabilistic prediction - predicts class membership probabilities based on Bayes Theorem Prior knowledge can be incorporated Easy implementation, Good results Dependence among attributes cannot be modeled – Bayesian Belief networks • Neural Networks Given enough hidden units and enough training samples, they can closely approximate any function Long training time Require a number of parameters typically best determined empirically Ability to classify untrained patterns Well-suited for continuous-valued inputs and outputs
- 21. 21 Machine Learning Techniques • Lazy Learning : K- Nearest Neighbor • Gaussian mixture model GMM) • Adapted GMM • Hidden Markov model (HMM) • Support vector Machine (SVM) • Posterior probability SVM
- 22. Supervised learning Supervised learning, as the name indicates, has the presence of a supervisor as a teacher. Basically supervised learning is when we teach or train the machine using data that is well labeled. Which means some data is already tagged with the correct answer. After that, the machine is provided with a new set of examples(data) so that the supervised learning algorithm analyses the training data(set of training examples) and produces a correct outcome from labeled data. 22
- 23. Training data 23
- 24. Testing data 24
- 26. Types of Supervised Learning Algorithms Supervised learning is classified into two categories of algorithms: Classification: A classification problem is when the output variable is a category, such as “Red” or “blue” or “disease” and “no disease”. Regression: A regression problem is when the output variable is a real value, such as “dollars” or “weight”. 26
- 27. Other types Supervised learning deals with or learns with “labeled” data. This implies that some data is already tagged with the correct answer. Other types:- Logistic Regression Naive Bayes Classifiers K-NN (k nearest neighbors) Decision Trees Support Vector Machine 27
- 28. Advantages:- Supervised learning allows collecting data and produces data output from previous experiences. Helps to optimize performance criteria with the help of experience. Supervised machine learning helps to solve various types of real-world computation problems. Disadvantages:- Classifying big data can be challenging. Training for supervised learning needs a lot of computation time. So, it requires a lot of time. 28
- 29. Unsupervised learning Unsupervised learning is the training of a machine using information that is neither classified nor labeled and allowing the algorithm to act on that information without guidance. Here the task of the machine is to group unsorted information according to similarities, patterns, and differences without any prior training of data. Unlike supervised learning, no teacher is provided that means no training will be given to the machine. Therefore the machine is restricted to find the hidden structure in unlabeled data by itself. 29
- 30. For instance, suppose it is given an image having both dogs and cats which it has never seen. 30
- 31. The machine has no idea about the features of dogs and cats so we can’t categorize it as ‘dogs and cats ‘. But it can categorize them according to their similarities, patterns, and differences, i.e., we can easily categorize the above picture into two parts. The first may contain all pics having dogs in them and the second part may contain all pics having cats in them. Here you didn’t learn anything before, which means no training data or examples. It allows the model to work on its own to discover patterns and information that was previously undetected. It mainly deals with unlabelled data. 31
- 32. Types of Unsupervised learning Unsupervised learning is classified into two categories of algorithms: Clustering: A clustering problem is where you want to discover the inherent groupings in the data, such as grouping customers by purchasing behavior. Association: An association rule learning problem is where you want to discover rules that describe large portions of your data, such as people that buy X also tend to buy Y 32
- 33. Other Types Types of Unsupervised Learning:- Clustering Exclusive (partitioning) Agglomerative Overlapping Probabilistic Clustering Types:- Hierarchical clustering K-means clustering Principal Component Analysis Singular Value Decomposition Independent Component Analysis 33
- 34. Parameters Supervised machine learning Unsupervised machine learning Input Data Algorithms are trained using labeled data. Algorithms are used against data that is not labeled Computational Complexity Simpler method Computationally complex Accuracy Highly accurate Less accurate 34
- 35. Reinforcement Learning Refer the below website: https://www.geeksforgeeks.org/what-is- reinforcement-learning/ 35