Learn to instruct a computer to acquire concepts using data, without being explicitly programmed. Acquire knowledge of the main methods of supervised and non-supervised machine learning, and discuss the properties and criteria of applicability. Acquire the ability to formulate correctly the problem, to choose the appropriate algorithm, and to perform the experimental analysis in order to evaluate the results obtained. Take care of the practical aspect of the implementation of the introduced methods by presenting different examples of use in different application scenarios.
Curriculum
teacher profile teaching materials
Supervised and unsupervised learning; Model representation; The cost function; The gradient descent algorithm;
Linear regression; The gradient descent for linear regression; Logistic regression; The gradient descent for logistic regression; The normal equation;
The problem of classification; The representation of the hypothesis; The cost function; The one-vs-all method; The problem of overfitting; Regularization in linear and logistic regression;
The perceptron; Le Neural Networks; The Error-back propagation algorithm; Random initialization of weights;
Model selection; The train, validation and test set; Diagnosis by bias and variance; The learning curves; Error analysis;
Support Vector Machines;
K-means clustering;
Principal Components Analysis for dimensionality reduction;
Anomaly Detection algorithms;
Recommender Systems;
Large scale machine learning systems including parallelized and map-reduce systems;
R.O. Duda, P.E. Hart, D.G. Stork. Pattern Classification (2001) John Wiley & Sons.
Fruizione: 20410432 IN550 – MACHINE LEARNING in Scienze Computazionali LM-40 CASTIGLIONE Filippo
Programme
Introduction to Machine Learning: what is machine learning; what is it aimed at, what are the problems; what are the theoretical tools used; overview of the topics that will be covered during the course.Supervised and unsupervised learning; Model representation; The cost function; The gradient descent algorithm;
Linear regression; The gradient descent for linear regression; Logistic regression; The gradient descent for logistic regression; The normal equation;
The problem of classification; The representation of the hypothesis; The cost function; The one-vs-all method; The problem of overfitting; Regularization in linear and logistic regression;
The perceptron; Le Neural Networks; The Error-back propagation algorithm; Random initialization of weights;
Model selection; The train, validation and test set; Diagnosis by bias and variance; The learning curves; Error analysis;
Support Vector Machines;
K-means clustering;
Principal Components Analysis for dimensionality reduction;
Anomaly Detection algorithms;
Recommender Systems;
Large scale machine learning systems including parallelized and map-reduce systems;
Core Documentation
C.M. Bishop, Pattern Recognition and Machine Learning, Springer, 2006R.O. Duda, P.E. Hart, D.G. Stork. Pattern Classification (2001) John Wiley & Sons.
Type of delivery of the course
The lessons will be held in the classroom using projections and the blackboard. The slides projected are the main study material.Attendance
Attendance is not compulsory but strongly recommended as part of the lessons will be done in the lab and will consist in coding pieces of machine learning algorithms seen in theory.Type of evaluation
Learning verification takes place through a written test lasting about 2 hours and an oral part. The written test is organized through a number of multiple choice questions and some other open exercises, aimed at verifying the level of effective understanding of the concepts and the students' ability to apply them in real contexts. teacher profile teaching materials
Supervised and unsupervised learning; Model representation; The cost function; The gradient descent algorithm;
Linear regression; The gradient descent for linear regression; Logistic regression; The gradient descent for logistic regression; The normal equation;
The problem of classification; The representation of the hypothesis; The cost function; The one-vs-all method; The problem of overfitting; Regularization in linear and logistic regression;
The perceptron; Le Neural Networks; The Error-back propagation algorithm; Random initialization of weights;
Model selection; The train, validation and test set; Diagnosis by bias and variance; The learning curves; Error analysis;
Support Vector Machines;
K-means clustering;
Principal Components Analysis for dimensionality reduction;
Anomaly Detection algorithms;
Recommender Systems;
Large scale machine learning systems including parallelized and map-reduce systems;
R.O. Duda, P.E. Hart, D.G. Stork. Pattern Classification (2001) John Wiley & Sons.
Fruizione: 20410432 IN550 – MACHINE LEARNING in Scienze Computazionali LM-40 CASTIGLIONE Filippo
Programme
Introduction to Machine Learning: what is machine learning; what is it aimed at, what are the problems; what are the theoretical tools used; overview of the topics that will be covered during the course.Supervised and unsupervised learning; Model representation; The cost function; The gradient descent algorithm;
Linear regression; The gradient descent for linear regression; Logistic regression; The gradient descent for logistic regression; The normal equation;
The problem of classification; The representation of the hypothesis; The cost function; The one-vs-all method; The problem of overfitting; Regularization in linear and logistic regression;
The perceptron; Le Neural Networks; The Error-back propagation algorithm; Random initialization of weights;
Model selection; The train, validation and test set; Diagnosis by bias and variance; The learning curves; Error analysis;
Support Vector Machines;
K-means clustering;
Principal Components Analysis for dimensionality reduction;
Anomaly Detection algorithms;
Recommender Systems;
Large scale machine learning systems including parallelized and map-reduce systems;
Core Documentation
C.M. Bishop, Pattern Recognition and Machine Learning, Springer, 2006R.O. Duda, P.E. Hart, D.G. Stork. Pattern Classification (2001) John Wiley & Sons.
Type of delivery of the course
The lessons will be held in the classroom using projections and the blackboard. The slides projected are the main study material.Attendance
Attendance is not compulsory but strongly recommended as part of the lessons will be done in the lab and will consist in coding pieces of machine learning algorithms seen in theory.Type of evaluation
Learning verification takes place through a written test lasting about 2 hours and an oral part. The written test is organized through a number of multiple choice questions and some other open exercises, aimed at verifying the level of effective understanding of the concepts and the students' ability to apply them in real contexts. teacher profile teaching materials
Supervised and unsupervised learning; Model representation; The cost function; The gradient descent algorithm;
Linear regression; The gradient descent for linear regression; Logistic regression; The gradient descent for logistic regression; The normal equation;
The problem of classification; The representation of the hypothesis; The cost function; The one-vs-all method; The problem of overfitting; Regularization in linear and logistic regression;
The perceptron; Le Neural Networks; The Error-back propagation algorithm; Random initialization of weights;
Model selection; The train, validation and test set; Diagnosis by bias and variance; The learning curves; Error analysis;
Support Vector Machines;
K-means clustering;
Principal Components Analysis for dimensionality reduction;
Anomaly Detection algorithms;
Recommender Systems;
Large scale machine learning systems including parallelized and map-reduce systems;
R.O. Duda, P.E. Hart, D.G. Stork. Pattern Classification (2001) John Wiley & Sons.
Fruizione: 20410432 IN550 – MACHINE LEARNING in Scienze Computazionali LM-40 CASTIGLIONE Filippo
Programme
Introduction to Machine Learning: what is machine learning; what is it aimed at, what are the problems; what are the theoretical tools used; overview of the topics that will be covered during the course.Supervised and unsupervised learning; Model representation; The cost function; The gradient descent algorithm;
Linear regression; The gradient descent for linear regression; Logistic regression; The gradient descent for logistic regression; The normal equation;
The problem of classification; The representation of the hypothesis; The cost function; The one-vs-all method; The problem of overfitting; Regularization in linear and logistic regression;
The perceptron; Le Neural Networks; The Error-back propagation algorithm; Random initialization of weights;
Model selection; The train, validation and test set; Diagnosis by bias and variance; The learning curves; Error analysis;
Support Vector Machines;
K-means clustering;
Principal Components Analysis for dimensionality reduction;
Anomaly Detection algorithms;
Recommender Systems;
Large scale machine learning systems including parallelized and map-reduce systems;
Core Documentation
C.M. Bishop, Pattern Recognition and Machine Learning, Springer, 2006R.O. Duda, P.E. Hart, D.G. Stork. Pattern Classification (2001) John Wiley & Sons.
Type of delivery of the course
The lessons will be held in the classroom using projections and the blackboard. The slides projected are the main study material.Attendance
Attendance is not compulsory but strongly recommended as part of the lessons will be done in the lab and will consist in coding pieces of machine learning algorithms seen in theory.Type of evaluation
Learning verification takes place through a written test lasting about 2 hours and an oral part. The written test is organized through a number of multiple choice questions and some other open exercises, aimed at verifying the level of effective understanding of the concepts and the students' ability to apply them in real contexts.