Servizi per la didattica
PORTALE DELLA DIDATTICA

Machine learning for pattern recognition

01SCSIU

A.A. 2018/19

2018/19

Machine learning for pattern recognition

PERIOD: MARCH This course gives a broad yet rigorous introduction to machine learning and statistical pattern recognition. It focuses on supervised generative and discriminative learning models, analyzing some important topics such as model architectures, training and evaluation techniques. The course will compare different models and classification approaches on the popular MNIST digit recognition dataset. It will also discuss about the applications of the proposed machine learning approaches to image, speech, and speaker recognition. Students will be asked to apply the acquired knowledge to develop their own classification system using labeled training and evaluation data provided during the course. Each system will then be evaluated on another unlabeled and previously unseen data set. This course belongs to an educational path on Data Science. The path is composed by - an introductory course (Data Mining: Concepts and Algorithms), covering data analytics fundamentals, which is a cultural prerequisite for the other courses - 5 thematic courses dealing in depth with specific Data Science topics, such as different algorithm types or application domains: - Data Analytics for Science and Society - Machine Learning for Pattern Recognition - Mimetic Learning - Text Mining and Analytics - Visualization and Visual Analytics

Machine learning for pattern recognition

PERIOD: MARCH This course gives a broad yet rigorous introduction to machine learning and statistical pattern recognition. It focuses on supervised generative and discriminative learning models, analyzing some important topics such as model architectures, training and evaluation techniques. The course will compare different models and classification approaches on the popular MNIST digit recognition dataset. It will also discuss about the applications of the proposed machine learning approaches to image, speech, and speaker recognition. Students will be asked to apply the acquired knowledge to develop their own classification system using labeled training and evaluation data provided during the course. Each system will then be evaluated on another unlabeled and previously unseen data set. This course belongs to an educational path on Data Science. The path is composed by - an introductory course (Data Mining: Concepts and Algorithms), covering data analytics fundamentals, which is a cultural prerequisite for the other courses - 5 thematic courses dealing in depth with specific Data Science topics, such as different algorithm types or application domains: - Data Analytics for Science and Society - Machine Learning for Pattern Recognition - Mimetic Learning - Text Mining and Analytics - Visualization and Visual Analytics

Machine learning for pattern recognition

Machine learning for pattern recognition

Machine learning for pattern recognition

Machine learning for pattern recognition

Machine learning for pattern recognition

- Introduction to pattern classification - Decision theory - Classification model taxonomy - Generative probabilistic models - Discriminative probabilistic models - Decision functions - Model Evaluation - Binary classification - Accuracy measures - ROC / DET curves - Multiclass classification - Generative probabilistic models - Gaussian models - Gaussian mixture models - Expectation-Maximization for GMMs - Time-dependent models: Hidden Markov Models - Inference with HMMs - EM for HMMs - Latent variable models: Factor Analysis - Introduction to Factor Analysis - Linear-Gaussian models - Probabilistic Principal Component Analysis - Probabilistic Linear Discriminant Analysis - Subspace models for GMMs

Machine learning for pattern recognition

- Introduction to pattern classification - Decision theory - Classification model taxonomy - Generative probabilistic models - Discriminative probabilistic models - Decision functions - Model Evaluation - Binary classification - Accuracy measures - ROC / DET curves - Multiclass classification - Generative probabilistic models - Gaussian models - Gaussian mixture models - Expectation-Maximization for GMMs - Time-dependent models: Hidden Markov Models - Inference with HMMs - EM for HMMs - Latent variable models: Factor Analysis - Introduction to Factor Analysis - Linear-Gaussian models - Probabilistic Principal Component Analysis - Probabilistic Linear Discriminant Analysis - Subspace models for GMMs

Machine learning for pattern recognition

Schedule: - Friday 08/03/2019 from 15:00 to 17:00 ROOM C - Monday 11/03/2019 from 15:00 to 18:00 ROOM C - Wednesday 13/03/2019 from 15:00 to 18:00 ROOM C - Friday 15/03/2019 from 15:00 to 18:00 ROOM C - Monday 18/03/2019 from 15:00 to 6:00 pm ROOM C - Wednesday 20/03/2019 from 15:00 to 18:00 ROOM C - Friday 22/03/2019 from 15:00 to 18:00 ROOM C

Machine learning for pattern recognition

Schedule: - Friday 08/03/2019 from 15:00 to 17:00 ROOM C - Monday 11/03/2019 from 15:00 to 18:00 ROOM C - Wednesday 13/03/2019 from 15:00 to 18:00 ROOM C - Friday 15/03/2019 from 15:00 to 18:00 ROOM C - Monday 18/03/2019 from 15:00 to 6:00 pm ROOM C - Wednesday 20/03/2019 from 15:00 to 18:00 ROOM C - Friday 22/03/2019 from 15:00 to 18:00 ROOM C

Machine learning for pattern recognition

Machine learning for pattern recognition

Machine learning for pattern recognition

Machine learning for pattern recognition

Machine learning for pattern recognition

ModalitÓ di esame:

Machine learning for pattern recognition

Machine learning for pattern recognition

Exam:

Machine learning for pattern recognition



© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY
m@il