Probabilistic Machine Learning

Recommended Prerequisites

Probabilities and Statistics, Pattern Recognition, Introduction to Machine Learning 

Teaching Methods

The lectures (T) will cover key concepts, theories, techniques and algorithms related to probabilistic machine learning, Bayesian inference and regression. The practical lectures (PL), an important part of the module, will provide opportunities to consolidate what has been learned during the lecturers (T). The practical work (ie, coursework) will be done under the supervision of the Lecturer and/or teaching assistant(s). Marking (ie, grading) will be based essentialy on 3 components: (i) a short report or presentation; (ii) concise project related to a practical problem; (iii) written exam.

Learning Outcomes

The aim of this module is for the students to understand the fundamentals of probabilistic machine learning (ML) and their applications in robotics and automation engineering problems. This module will introduce and discuss modern learning techniques, random variables, uncertainty, algorithms, time-varying problems, and Bayesian inference in a machine learning perspective. Content wise, this module will cover: probabilistic inference, Bayesian Networks (BN), Dynamic Bayesian Networks (DBN), ML applied to regression (dynamic) problems, decision-making, advanced sensor/data fusion.

By the end of the module, student will have a significant understanding and a practical experience on how to formulate and solve real problems and will master the most recent techniques of machine learning as well. He/she will be capable of designing, implementing, testing and validating solutions for real-world engineering problems related to robotics, autonomous systems and automation, based on a probabi

Work Placement(s)

No

Syllabus

1. Introduction: probabilistic ML applied to robotics, autonomous driving, dynamic systems, and automation

2. Probability and stochastic processes

3. Fundamentals of cost functions

4. Linear and non-linear regression (ML perspective)

5. Probabilistic models, Bayesian inference

6. Bayesian Networks (BN)

7. Dynamic BN (DBN)

8. Probabilistic machine learning

9. Applications, ROS, case studies related to robotics, autonomous driving, automation, combining methods     

Head Lecturer(s)

Cristiano Premebida

Assessment Methods

Assessment
Report of a seminar or field trip: 10.0%
Project: 40.0%
Exam: 50.0%

Bibliography

Kevin P. Murphy "Machine Learning: a Probabilistic Perspective", the MIT Press (2012)

A. Papoulis, S.U. Pillai. “Probability, Random Variables and Stochastic Processes”, (2002)

Sebastian Thrun, Wolfram Burgard, Dieter Fox. “Probabilistic robotics”, MIT Press, (2006)

Jerome H. Friedman, Robert Tibshirani, and Trevor Hastie. "The Elements of Statistical Learning", (2009)

David Barber "Bayesian Reasoning and Machine Learning", Cambridge University Press (2012)

Ian Goodfellow, Yoshua Bengio and Aaron Courville, "Deep Learning", MIT Press (2016)

Christopher M. Bishop "Pattern Recognition and Machine Learning". Springer (2006)

Richard S. Sutton and Andrew G. Barto. "Reinforcement Learning, An Introduction" (2018)