Univ.-Prof. Dr. Elmar Rueckert was teaching this course at the University of Luebeck in the winter semester in 2018.
Language: English only
Course Details
[WS2018/19] In the winter semester, I will teach a course on Probabilistic Learning for Robotics which covers advanced topics including graphical models, factor graphs, probabilistic inference for decision making and planning, and computational models for inference in neuroscience. The lecture will take place in the Seminarraum Informatik 5 (Von Neumann) 2.132 from 12.00 – 14.00 on selected Thursdays.
In accompanying exercises and hands on tutorials the students will experiment with state of the art machine learning methods and robotic simulation tools. In particular, Mathworks’ MATLAB, the robot middleware ROS and the simulation tool V-Rep will be used. The exercises and tutorials will also take place in the seminar room 2.132 on selected Fridays (see the course materials and dates below).
Univ.-Prof. Dr. Elmar Rueckert was teaching this course at the University of Luebeck in 2018, 2019 and 2020.
Language: English only
Course Details
In this research seminar we discuss state of the art research topics in robotics, machine learning and autonomous systems. Presenters are invited guest speakers, researcher, and graduate and under graduate students.
Some remarks on the UzL Module idea: The lecture Probabilistic Machine Learning belongs to the Module Robot Learning (RO4100). In the winter semester, Prof. Dr. Elmar Rueckert is teaching the course Probabilistic Machine Learning (RO5101 T). In the summer semester, Prof. Dr. Elmar Rueckert is teaching the course Reinforcement Learning (RO5102 T). Important: Due to the study regulations, students have to attend both lectures to receive a final grade. Thus, there will be only a single written exam for both lectures. You can register for the written exam at the end of a semester.
Important dates
Written exam: 04. February 2021 (2nd appointment 04.03.2021)
Assignment I: Freitag, 11. Dezember 2020, 23:00
Assignment II: Freitag, 22. Januar 2021, 23:00
The course topics are
Introduction to Probability Theory (Statistics refresher, Bayes Theorem, Common Probability distributions, Gaussian Calculus).
Linear Probabilistic Regression (Linear models, Maximum Likelihood, Bayes & Logistic Regression).
Nonlinear Probabilistic Regression (Radial basis function networks, Gaussian Processes, Recent research results in Robotic Movement Primitives, Hierarchical Bayesian & Mixture Models).
Probabilistic Inference for Filtering, Smoothing and Planning (Classic, Extended & Unscented Kalman Filters, Particle Filters, Gibbs Sampling, Recent research results in Neural Planning).
Q & A Session on Mondays ,12:15-13:45, virtual using the WEBEX room of Nils Rottmann. The session will be closed if no questions where asked till 12:20.
Interactive Online Lectures
In the lecture, Prof. Rueckert is using a self made lightboard to ensure an interactive and professional teaching environment. Have a look at the post on how to build such a lightboard. Here is an example recording.
Requirements
Strong statistical and mathematical knowledge is required beforehand. It is highly recommended to attend the course Humanoid Robotics (RO5300) prior to attending this course. The students will also experiment with state-of-the-art machine learning methods and robotic simulation tools which require strong programming skills.
Grading
The course is accompanied by twowritten assignments. Both assignments have to be passed as requirement to attend the written exam. Details will be presented in the first course unit on October the 22nd, 2020.
Materials for the Exercise
The course is accompanied by three graded assignments on Probabilistic Regression, Probabilistic Inference and on Probabilistic Optimization. The assignments will include algorithmic implementations in Matlab, Python or C++ and will be presented during the exercise sessions. The Robot Operating System (ROS) will also be part in some assignments as well as the simulation environment Gazebo. To experiment with state-of-the-art robot control and learning methods Mathworks’ MATLAB will be used. If you do not have it installed yet, please follow the instructions of our IT-Service Center.
Literature
Daphne Koller, Nir Friedman. Probabilistic Graphical Models: Principles and Techniques. ISBN 978-0-262-01319-2
Christopher M. Bishop. Pattern Recognition and Machine Learning. Springer (2006). ISBN 978-0-387-31073-2.
David Barber. Bayesian Reasoning and Machine Learning, Cambridge University Press (2012). ISBN 978-0-521-51814-7.
Kevin P. Murphy. Machine Learning: A Probabilistic Perspective. ISBN 978-0-262-01802-9
The lecture Reinforcement Learning belongs to the Module Robot Learning (RO4100).
In the winter semester, Prof. Dr. Elmar Rueckert is teaching the course Probabilistic Machine Learning – PML (RO5101 T).
In the summer semester, Prof. Dr. Elmar Rueckert is teaching the course Reinforcement Learning – RL (RO4100 T).
Important Remarks
Students will receive a single grade for the Module Robot Learning (RO4100) based on the average grade of PML and RL (rounded down in favor of the students).
This course is organized through online lectures and exercises. Details to the organizations will be discussed in our
FIRST MEETING: 17.04.2020 12:15-13:45
using the WEBEX tool. Please follow the instructions of the ITSC here to setup your computer. Click on the links to create a google calendar event, joint the WEBEX meeting or to access the online slides.
Dates & Times of the Online Webex Meetings
Lectures are organized on FRIDAYS, 12:15-13:45, WEBEX Link
Exercises are organized on THURSDAYS, 09:15-10:00, WEBEX Link
Course description
Introduction to Robotics and Reinforcement Learning (Refresher on Robotics, kinematics, model learning and learning feedback control strategies).
Foundations of Decision Making (Reward Hypothesis, Markov Property, Markov Reward Process, Value Iteration, Markov Decision Process, Policy Iteration, Bellman Equation, Link to Optimal Control).
Principles of Reinforcement Learning (Exploration and Exploitation strategies, On & Off-policy learning, model-free and model-based policy learning, Algorithmic principles: Q-Learning, SARSA, (Multi-step) TD-Learning, Eligibility Traces).
Deep Reinforcement Learning (Introduction to Deep Networks, Stochastic Gradient Descent, Function Approximation, Fitted Q-Iteration, (Double) Deep Q-Learning, Policy-Gradient approaches, Recent research results in Stochastic Deep Neural Networks).
The learning objectives / qualifications are
Students get a comprehensive understanding of basic decision making theories, assumptions and methods.
Students learn to analyze the challenges in a reinforcement learning application and to identify promising learning approaches.
Students will understand the difference between deterministic and probabilistic policies and can define underlying assumptions and requirements for learning them.
Students understand and can apply advanced policy gradient methods to real world problems.
Students know how to analyze the learning results and improve the policy learner parameters.
Students understand how the basic concepts are used in current state of the art research in robot reinforcement learning and in deep neural networks.
Basic knowledge in Machine Learning and Neural Networks is required. It is highly recommended to attend any of (but not restricted to) the following courses Probabilistic Machine Learning (RO 5101 T), Artificial Intelligence II (CS 5204 T), Machine Learning (CS 5450), Medical Deep Learning (CS 4374) prior to attending this course. The students will also experiment with state-of-the-art Reinforcement Learning (RL) methods on benchmark RL simulator (OpenAI Gym, Pybullet), which requires strong Python programming skills and knowledge on Pytorch is preferred. All assignment related materials have been tested on a windows machine (Win10 platform).
Grading
The course grades will be computed solely from submitted student reports of six assignments. The reports and the code have to be submitted (one report per team) to xue@rob.uni-luebeck.de. Please note the list of dates and deadlines below. Each assignment has minimally two-week deadline, some of them are of longer duration. Please use Latex for writing your report.
Bonus Points
tudents can get Bonus Points (BP) during the lectures when all quiz questions are correctly answered (1 BP per lecture). In the assignments, BPs will be given to the students when optional (and often also challenging) tasks are implemented and discussed.
Materials for the Exercise
The course is accompanied by pieces of course work on policy search for discrete state and action spaces (grid world example), policy learning in continuous spaces using function approximations and policy gradient methods in challenging simulated robotic tasks. The theoretical assignment questions are based on the lecture and also on the first three literature sources listed above. It is strongly recommended to read (or watch) these material in parallel to attending lecture. The assignments will include both written tasks and algorithmic implementations in Python. The tasks will be presented during the exercise sessions. As simulation environment, the OpenAI Gym platform will be used in the project works.
Literature
Richard S. Sutton, Andrew Barto: Reinforcement Learning: An Introductionsecond edition. The MIT Press Cambridge, Massachusetts London, England, 2018. Link to the online book (PDF)
Csaba Szepesvri: Algorithms for Reinforcement Learning. Morgan & Claypool in July 2010.
B. Siciliano, L. Sciavicco: Robotics: Modelling,Planning and Control, Springer, 2009.
Puterman, Martin L. Markov decision processes: discrete stochastic dynamic programming. John Wiley & Sons, 2014.
Szepesvari, Csaba. Algorithms for reinforcement learning (synthesis lectures on artificial intelligence and machine learning). Morgan and Claypool (2010).
using the webex tool. Please follow the instructions of the ITSC here to setup your computer. Click on the links to create a google calendar event, joint the webex meeting or to access the online slides.
during a webex meeting. The slides are available here. Click on the links to create a google calendar event, joint the webex meeting or to access the online slides.
Dates & Times of the Online Webex Meetings
Lectures are organized on TUESDAYS, 10:15-11:45, Webex Link
Exercises are organized on WEDNESDAYS, 10:15-11:45, Webex Link
Course Description
Prof. Dr. Elmar Rueckert is teaching the course on Humanoid Robotics (RO5300) together with M.Sc. Nils Rottmann, who supervises the exercises. In this course he discusses the key components of one of the most complex autonomous systems. These topics are
This course provides a unique overview over central topics in robotics. A particular focus is put in the dependencies and interaction among the components in the control loop illustrated in the image above. These interactions are discussed in the context of state of the art methods including dynamical systems movement primitives, gradient based policy search methods or probabilisitic inference for planning algorithms.
In sum, the lecture provides a structured and well motivated overview over modern techniques and tools which enable the students to define reward functions, implement robot controller and interaction software and to apply and extend state of the art reinforcement learning and planning approaches.
No special knowledge is required beforehand. All concepts and theories will be developed during the lectures or the tutorials.
The students will also experiment with state of the art machine learning methods and robotic simulation tools in accompanying exercises. Hands on tutorials on programming with Matlab and the simulation tool V-Rep complement the course content.
Grading
The course grades will be computed solely from submitted student reports of fourassignments, for each you can get 25 points. Two weeks after the assignment presentation events, the reports and the code have to be submitted (one report per team) to hum_rob@rob.uni-luebeck.de. Below is the list of dates and deadlines. Please use Latex for submitting the assignments.
Bonus Points
You can receive up to 30Bonus Points (BP) during the course, 10 BP during the lectures and 20 BP for submitting optional exercise solutions. To get BPs during the lecture, you have to successfully participate at the quizz sessions at the beginning of each lecture. To get BPs for the optional exercise solutions, you have to (clearly and readable) write down your solution, take a photo and send it to hum_rob@rob.uni-luebeck.de with the concern Exercise_##_LastName, where ## is the number of the exercise. You have to send your solution prior to the start of the exercise session where we will cover the exercise sheet.
Points to Grades
>= Points
Grade
Comment
95
1.0
Best possible grade
90
1.3
85
1.7
80
2.0
75
2.3
70
2.7
65
3.0
60
3.3
55
3.7
50
4.0
0
5.0
Worst possible grade
Materials for the Exercise
For simulating robot manipulation tasks we will use the simulator V-REP. For research and for teaching a free eduction version can be found here. To experiment with state of the art robot control and learning methods Mathworks’ MATLAB will be used. If you do not have it installed yet, please follow the instructions of our IT-Service Center.
This lecture provides a unique overview over central topics in Cyber-Physical-Systems:
Kinematics, Dynamics & Simulation of CPS
Data Representations & Model Learning
Feedback Control, Priorities & Torque Control
Planning & Cognitive Reasoning
Reinforcement Learning & Policy Search
The course provides a structured and well motivated overview over modern techniques and tools which enable the students to define learning problems in Cyber-Physical-Systems.
Students get a comprehensive understanding of Cyber-Physical-Systems.
Students learn to analyze the challenges in simulating, modeling and controlling CPS.
Students understand and can apply basic machine learning and control techniques in CPS.
Students know how to analyze the models’ results, improve the model parameters and can interpret the model predictions and their relevance.
Programming Assignments & Simulation Tools
For simulating robotic systems, we will use the tool CoppeliaSim. The tool can be used for free for research and for teaching.
To experiment with state of the art robot control and learning methods Python will be used. If you never used Python and are unexperienced in programming, please visit the tutorials on Python programming prior to the lecture.
The course will also use the tool Code With Me from JetBrains. With this stool, we can develop jointly code.
Literature
The Probabilistic Machine Learning book by Univ.-Prof. Dr. Elmar Rueckert.
Bishop 2006. Pattern Recognition and Machine Learning, Springer.
Barber 2007. Bayesian Reasoning and Machine Learning, Cambridge University Press.
Murray, Li and Sastry 1994. A mathematical introduction to robotic manipulation, CRC Press.
B. Siciliano, L. Sciavicco 2009. Robotics: Modelling,Planning and Control, Springer.
Kevin M. Lynch and Frank C. Park 2017. MODERN ROBOTICS, MECHANICS, PLANNING, AND CONTROL, Cambridge University Press.
Dates: Fridays13:15 – 14:45 Attention:There is one exception, on the 08.03.24, we start at 12:15!
Learning objectives / qualifications
Hands-on experience with machine learning methods.
Course Resources
Everybody needs to sign in at MUOnline for the course. Access to Moodle will be essential for all participants in the course, as it will serve as the platform where course materials will be made available.
It is written for experienced undergraduates or for first semester graduate students.
The lecture provides the basic knowledge for the application of modern machine learning methods. It includes an introduction to the basics of data modeling and probability theory. Classical probabilistic linear and non-linear regression methods are derived and discussed using practical examples.
26.06.2024 Exam Results & Best Practices & Feedback Discussion
Course Topics
Introduction to Machine Learning (Data and modelling fundamentals)
Introduction to Probability Theory (Statistics refresher, Bayes Theorem, Common Probability distributions, Gaussian Calculus).
Linear Probabilistic Regression (Linear models, Maximum Likelihood, Bayes & Logistic Regression).
Nonlinear Probabilistic Regression (Radial basis function networks, Gaussian Processes, Recent research results in Robotic Movement Primitives, Hierarchical Bayesian & Mixture Models).
Probabilistic Inference for Time Series (Time series data, basis function models, learning).
Learning objectives / qualifications
Students get a comprehensive understanding of basic probability theory concepts and methods.
Students learn to analyze the challenges in a task and to identify promising machine learning approaches.
Students will understand the difference between deterministic and probabilistic algorithms and can define underlying assumptions and requirements.
Students understand and can apply advanced regression, inference and optimization techniques to real world problems.
Students know how to analyze the models’ results, improve the model parameters and can interpret the model predictions and their relevance.
Students understand how the basic concepts are used in current state-of-the-art research in robot movement primitive learning and in neural planning.
Grading
The course will be graded based on a written exam (100 Points). 50% of all questions need to be answered correctly to be positive. The exam will take place in the classroom, or online, depending on the current university regulations.
In addition, up to 10 bonus points obtained in regular quiz sessions in the classroom, and 20% of the achieved points of the Machine Learning Lab will be added to your exam result. Note that bonus points can only be obtained when attending the lectures in person.