# Search result: Catalogue data in Spring Semester 2020

Mechanical Engineering Master | ||||||

Core Courses | ||||||

Robotics, Systems and Control The courses listed in this category “Core Courses” are recommended. Alternative courses can be chosen in agreement with the tutor. . | ||||||

Number | Title | Type | ECTS | Hours | Lecturers | |
---|---|---|---|---|---|---|

151-0310-00L | Model Predictive Engine Control Number of participants limited to 55. | W | 4 credits | 2V + 1U | T. Albin Rajasingham | |

Abstract | For efficient and stable operation of an internal combustion engine a multitude of complex control tasks have to be handled. In this lecture the application of model predictive control for these control challenges is introduced. | |||||

Objective | - Learn how to design and implement model predictive control algorithms for the example system “combustion engine”. Get to know the entire process from simulation-based control development to the application at a real-world combustion engine. - Deepen the knowledge concerning the necessary control algorithms for a combustion engine. | |||||

Content | - Physical phenomena and models for processes of the combustion engine such as air path and fuel path - Analysis of the control tasks arising in engine systems - Case studies for the application of model predictive control for combustion engines with the goal to handle the complex, multivariable system dynamics - Fundamentals of the implementation of model predictive control | |||||

Lecture notes | Lecture slides will be provided after each lecture. | |||||

Literature | L. Guzzella / C. Onder: "Introduction to Modeling and Control of Internal Combustion Engine Systems", J. Maciejowski: "Predictive Control with Constraints" | |||||

Prerequisites / Notice | Engine Systems (recommended). | |||||

151-0314-00L | Information Technologies in the Digital Product | W | 4 credits | 3G | E. Zwicker, R. Montau | |

Abstract | Objectives, Concepts and Methods of Digitalization, Digital Product and Product Lifecycle Management (PLM), Industry 4.0 Concepts for Digitalization: Product Structures, Optimization of Engineering Processes with digital models in Sales, Production, Service, Digital Twin versus Digital Thread PLM Fundamentals: Objects, Structures, Processes, Integrations, Visualization Best Practices | |||||

Objective | Students learn the fundamentals and concepts of Digitalization along the in the product lifecycle on the foundation of Product Lifecycle Management (PLM) technologies, the usage of databases, the integration of CAx systems and Visualization/AR, the configuration of computer-based collaboration leveraging IT-standards as well as variant and configuration management to enable an efficient utilization of the digital product approach in industry 4.0. | |||||

Content | Possibilities and potential of modern IT applications focussing on PLM and CAx technologies for targeted utilization in the context of product platform - business processes - IT tools. Introduction to the concepts of Product Lifecycle Management (PLM): information modeling, data management, revision, usage and distribution of product data. Structure and functional principles of PLM systems. Integration of new IT technologies in business processes. Possibilities of publication and automatic configuration of product variants via the Internet. Using state-of-the-art information and communication technologies to develop products globally across distributed locations. Interfaces in computer-integrated product development. Selection, configuration, adaptation and introduction of PLM systems. Examples and case studies for industrial usage of modern information technologies. Learning modules: - Introduction to Digitalization (Digital Product, PLM technology) - Database technology (foundation of digitalization) - Object Management - Object Classification - Object identification with Part Numbering Systems - CAx/PLM integration with Visualization/AR - Workflow & Change Management - Interfaces of the Digital Product - Enterprise Application Integration (EAI) | |||||

Lecture notes | Didactic concept / learning materials: The course consists of lectures and exercises based on practical examples. Provision of lecture handouts and script digitally in Moodle. | |||||

Prerequisites / Notice | Prerequisites: None Recommended: Fokus-Project, interest in Digitalization Lecture appropriate for D-MAVT, D-MTEC, D-ITET and D-INFK Testat/Credit Requirements / Exam: - execution of exercises in teams (recommended) - Oral exam 30 minutes, based on concrete problem cases | |||||

151-0318-00L | Ecodesign - Environmental-Oriented Product Development | W | 4 credits | 3G | R. Züst | |

Abstract | Ecodesign has a great potential to improve the environmental performance of a product. Main topics of the lecture: Motivation for Ecodesign; Methodical basics (defining environmental aspects; improvement strageies and measures); Ecodesign implementation (systematic guidance on integrating environmental considerations into product development) in a small project. | |||||

Objective | Experience shows that a significant part of the environmental impact of a business venture is caused by its own products in the pre and post-production areas. The goal of eco design is to reduce the total effect of a product on the environment in all phases of product life. The systematic derivation of promising improvement measures at the start of the product development process is a key skill that will be taught in the lectures. The participants will discover the economic and ecological potential of ECODESIGN and acquire competence in determining goal-oriented and promising improvements and will be able to apply the knowledge acquired on practical examples. | |||||

Content | Die Vorlesung ist in drei Blöcke unterteilt. Hier sollen die jeweiligen Fragen beantwortet werden: A) Motivation und Einstieg ins Thema: Welche Material- und Energieflüsse werden durch Produkte über alle Lebensphasen, d.h. von der Rohstoffgewinnung, Herstellung, Distribution, Nutzung und Entsorgungen verursacht? Welchen Einfluss hat die Produktentwicklung auf diese Auswirkungen? B) Grundlagen zum ECODESIGN PILOT: Wie können systematisch – über alle Produktlebensphasen hinweg betrachtet – bereits zu Beginn der Produktentwicklung bedeutende Umweltauswirkungen erkannt werden? Wie können zielgerichtet diejenigen Ecodesign-Maßnahmen ermittelt werden, die das größte ökonomische und ökologische Verbesserungspotential beinhalten? C) Anwendung des ECODESIGN PILOT: Welche Produktlebensphasen bewirken den größten Ressourcenverbrauch? Welche Verbesserungsmöglichkeiten bewirken einen möglichst großen ökonomischen und ökologischen Nutzen? Im Rahmen der Vorlesung werden verschiedene Praktische Beispiel bearbeitet. | |||||

Lecture notes | Für den Einstieg ins Thema ECODESIGN wurde verschiedene Lehrunterlagen entwickelt, die im Kurs zur Verfügung stehen und teilwesie auch ein "distance learning" ermöglichen: Lehrbuch: Wimmer W., Züst R.: ECODESIGN PILOT, Produkt-Innovations-, Lern- und Optimierungs-Tool für umweltgerechte Produktgestaltung mit deutsch/englischer CD-ROM; Zürich, Verlag Industrielle Organisation, 2001. ISBN 3-85743-707-3 CD: im Lehrbuch inbegriffen (oder Teil "Anwenden" on-line via: www.ecodesign.at) Internet: www.ecodesign.at vermittelt verschiedene weitere Zugänge zum Thema. Zudem werden CD's abgegeben, auf denen weitere Lehrmodule vorhanden sind. | |||||

Literature | Hinweise auf Literaturen werden on-line zur Verfügung gestellt. | |||||

Prerequisites / Notice | Testatbedingungen: Abgabe von zwei Übungen | |||||

151-0530-00L | Nonlinear Dynamics and Chaos II | W | 4 credits | 4G | G. Haller | |

Abstract | The internal structure of chaos; Hamiltonian dynamical systems; Normally hyperbolic invariant manifolds; Geometric singular perturbation theory; Finite-time dynamical systems | |||||

Objective | The course introduces the student to advanced, comtemporary concepts of nonlinear dynamical systems analysis. | |||||

Content | I. The internal structure of chaos: symbolic dynamics, Bernoulli shift map, sub-shifts of finite type; chaos is numerical iterations. II.Hamiltonian dynamical systems: conservation and recurrence, stability of fixed points, integrable systems, invariant tori, Liouville-Arnold-Jost Theorem, KAM theory. III. Normally hyperbolic invariant manifolds: Crash course on differentiable manifolds, existence, persistence, and smoothness, applications. IV. Geometric singular perturbation theory: slow manifolds and their stability, physical examples. V. Finite-time dynamical system; detecting Invariant manifolds and coherent structures in finite-time flows | |||||

Lecture notes | Students have to prepare their own lecture notes | |||||

Literature | Books will be recommended in class | |||||

Prerequisites / Notice | Nonlinear Dynamics I (151-0532-00) or equivalent | |||||

151-0534-00L | Advanced Dynamics | W | 4 credits | 3V + 1U | P. Tiso | |

Abstract | Lagrangian dynamics - Principle of virtual work and virtual power - holonomic and non holonomic contraints - 3D rigid body dynamics - equilibrium - linearization - stability - vibrations - frequency response | |||||

Objective | This course provides the students of mechanical engineering with fundamental analytical mechanics for the study of complex mechanical systems .We introduce the powerful techniques of principle of virtual work and virtual power to systematically write the equation of motion of arbitrary systems subjected to holonomic and non-holonomic constraints. The linearisation around equilibrium states is then presented, together with the concept of linearised stability. Linearized models allow the study of small amplitude vibrations for unforced and forced systems. For this, we introduce the concept of vibration modes and frequencies, modal superposition and modal truncation. The case of the vibration of light damped systems is discussed. The kinematics and dynamics of 3D rigid bodies is also extensively treated. | |||||

Lecture notes | Lecture notes are produced in class and are downloadable right after each lecture. | |||||

Literature | The students will prepare their own notes. A copy of the lecture notes will be available. | |||||

Prerequisites / Notice | Mechanics III or equivalent; Analysis I-II, or equivalent; Linear Algebra I-II, or equivalent. | |||||

151-0566-00L | Recursive Estimation | W | 4 credits | 2V + 1U | R. D'Andrea | |

Abstract | Estimation of the state of a dynamic system based on a model and observations in a computationally efficient way. | |||||

Objective | Learn the basic recursive estimation methods and their underlying principles. | |||||

Content | Introduction to state estimation; probability review; Bayes' theorem; Bayesian tracking; extracting estimates from probability distributions; Kalman filter; extended Kalman filter; particle filter; observer-based control and the separation principle. | |||||

Lecture notes | Lecture notes available on course website: http://www.idsc.ethz.ch/education/lectures/recursive-estimation.html | |||||

Prerequisites / Notice | Requirements: Introductory probability theory and matrix-vector algebra. | |||||

151-0630-00L | Nanorobotics | W | 4 credits | 2V + 1U | S. Pané Vidal | |

Abstract | Nanorobotics is an interdisciplinary field that includes topics from nanotechnology and robotics. The aim of this course is to expose students to the fundamental and essential aspects of this emerging field. | |||||

Objective | The aim of this course is to expose students to the fundamental and essential aspects of this emerging field. These topics include basic principles of nanorobotics, building parts for nanorobotic systems, powering and locomotion of nanorobots, manipulation, assembly and sensing using nanorobots, molecular motors, and nanorobotics for nanomedicine. | |||||

151-0634-00L | Perception and Learning for Robotics Number of participants limited to: 30 To apply for the course please create a CV in pdf of max. 2 pages, including your machine learning and/or robotics experience. Please send the pdf to cesarc@ethz.ch for approval. | W | 4 credits | 9A | C. D. Cadena Lerma, J. J. Chung | |

Abstract | This course covers tools from statistics and machine learning enabling the participants to deploy these algorithms as building blocks for perception pipelines on robotic tasks. All mathematical methods provided within the course will be discussed in context of and motivated by example applications mostly from robotics. The main focus of this course are student projects on robotics. | |||||

Objective | Applying Machine Learning methods for solving real-world robotics problems. | |||||

Content | Deep Learning for Perception; (Deep) Reinforcement Learning; Graph-Based Simultaneous Localization and Mapping | |||||

Lecture notes | Slides will be made available to the students. | |||||

Literature | Will be announced in the first lecture. | |||||

Prerequisites / Notice | The students are expected to be familiar with material of the "Recursive Estimation" and the "Introduction to Machine Learning" lectures. Particularly understanding of basic machine learning concepts, stochastic gradient descent for neural networks, reinforcement learning basics, and knowledge of Bayesian Filtering are required. Furtheremore, good knowledge of programming in C++ and Python is required. | |||||

151-0641-00L | Introduction to Robotics and Mechatronics Number of participants limited to 60. Enrollment is only valid through registration on the MSRL website (www.msrl.ethz.ch). Registrations per e-mail is no longer accepted! | W | 4 credits | 2V + 2U | B. Nelson, N. Shamsudhin | |

Abstract | The aim of this lecture is to expose students to the fundamentals of mechatronic and robotic systems. Over the course of these lectures, topics will include how to interface a computer with the real world, different types of sensors and their use, different types of actuators and their use. | |||||

Objective | An ever-increasing number of mechatronic systems are finding their way into our daily lives. Mechatronic systems synergistically combine computer science, electrical engineering, and mechanical engineering. Robotics systems can be viewed as a subset of mechatronics that focuses on sophisticated control of moving devices. The aim of this course is to practically and theoretically expose students to the fundamentals of mechatronic and robotic systems. Over the course of the semester, the lecture topics will include an overview of robotics, an introduction to different types of sensors and their use, the programming of microcontrollers and interfacing these embedded computers with the real world, signal filtering and processing, an introduction to different types of actuators and their use, an overview of computer vision, and forward and inverse kinematics. Throughout the course, students will periodically attend laboratory sessions and implement lessons learned during lectures on real mechatronic systems. By the end of the course, you will be able to independently choose, design and integrate these different building blocks into a working mechatronic system. | |||||

Content | The course consists of weekly lectures and lab sessions. The weekly topics are the following: 0. Course Introduction 1. C Programming 2. Sensors 3. Data Acquisition 4. Signal Processing 5. Digital Filtering 6. Actuators 7. Computer Vision and Kinematics 8. Modeling and Control 9. Review and Outlook The lecture schedule can be found on our course page on the MSRL website (www.msrl.ethz.ch) | |||||

Prerequisites / Notice | The students are expected to be familiar with C programming. | |||||

151-0660-00L | Model Predictive Control | W | 4 credits | 2V + 1U | M. Zeilinger | |

Abstract | Model predictive control is a flexible paradigm that defines the control law as an optimization problem, enabling the specification of time-domain objectives, high performance control of complex multivariable systems and the ability to explicitly enforce constraints on system behavior. This course provides an introduction to the theory and practice of MPC and covers advanced topics. | |||||

Objective | Design and implement Model Predictive Controllers (MPC) for various system classes to provide high performance controllers with desired properties (stability, tracking, robustness,..) for constrained systems. | |||||

Content | - Review of required optimal control theory - Basics on optimization - Receding-horizon control (MPC) for constrained linear systems - Theoretical properties of MPC: Constraint satisfaction and stability - Computation: Explicit and online MPC - Practical issues: Tracking and offset-free control of constrained systems, soft constraints - Robust MPC: Robust constraint satisfaction - Nonlinear MPC: Theory and computation - Hybrid MPC: Modeling hybrid systems and logic, mixed-integer optimization - Simulation-based project providing practical experience with MPC | |||||

Lecture notes | Script / lecture notes will be provided. | |||||

Prerequisites / Notice | One semester course on automatic control, Matlab, linear algebra. Courses on signals and systems and system modeling are recommended. Important concepts to start the course: State-space modeling, basic concepts of stability, linear quadratic regulation / unconstrained optimal control. Expected student activities: Participation in lectures, exercises and course project; homework (~2hrs/week). | |||||

151-0854-00L | Autonomous Mobile Robots | W | 5 credits | 4G | R. Siegwart, M. Chli, N. Lawrance | |

Abstract | The objective of this course is to provide the basics required to develop autonomous mobile robots and systems. Main emphasis is put on mobile robot locomotion and kinematics, environment perception, and probabilistic environment modeling, localizatoin, mapping and navigation. Theory will be deepened by exercises with small mobile robots and discussed accross application examples. | |||||

Objective | The objective of this course is to provide the basics required to develop autonomous mobile robots and systems. Main emphasis is put on mobile robot locomotion and kinematics, environment perception, and probabilistic environment modeling, localizatoin, mapping and navigation. | |||||

Lecture notes | This lecture is enhanced by around 30 small videos introducing the core topics, and multiple-choice questions for continuous self-evaluation. It is developed along the TORQUE (Tiny, Open-with-Restrictions courses focused on QUality and Effectiveness) concept, which is ETH's response to the popular MOOC (Massive Open Online Course) concept. | |||||

Literature | This lecture is based on the Textbook: Introduction to Autonomous Mobile Robots Roland Siegwart, Illah Nourbakhsh, Davide Scaramuzza, The MIT Press, Second Edition 2011, ISBN: 978-0262015356 | |||||

151-1115-00L | Aircraft Aerodynamics and Flight Mechanics | W | 4 credits | 3G | J. Wildi | |

Abstract | Equations of motion. Aircraft flight perfomance, flight envelope. Aircraft static stability and control, longituadinal and lateral stbility. Dynamic longitudinal and lateral stability. Flight test. Wind tunnel test. | |||||

Objective | - Knowledge of methods to solve flight mechanic problems - To be able to apply basic methods for flight performence calculation and stability investigations - Basic knowledge of flight and wind tunnel tests and test evaluation methods | |||||

Content | Equations of motion. Aircraft flight perfomance, flight envelope. Aircraft static stability and control, longituadinal and lateral stbility. Dynamic longitudinal and lateral stability. Flight testing. Wind tunnel testing. | |||||

Lecture notes | Ausgewählte Kapitel der Flugtechnik (J. Wildi) | |||||

Literature | Mc Cormick, B.W.: Aerodynamics, Aeronautics and Flight Mechanics (John Wiley and Sons), 1979 / 1995 Anderson, J: Fundamentals of Aerodynamics (McGraw-Hill Comp Inc), 2010 | |||||

Prerequisites / Notice | Recommended: Lecture 'Basics of Aircraft und Vehicle Aerodynamics' (FS) | |||||

151-0116-10L | High Performance Computing for Science and Engineering (HPCSE) for Engineers II | W | 4 credits | 4G | P. Koumoutsakos, S. M. Martin | |

Abstract | This course focuses on programming methods and tools for parallel computing on multi and many-core architectures. Emphasis will be placed on practical and computational aspects of Uncertainty Quantification and Propagation including the implementation of relevant algorithms on HPC architectures. | |||||

Objective | The course will teach - programming models and tools for multi and many-core architectures - fundamental concepts of Uncertainty Quantification and Propagation (UQ+P) for computational models of systems in Engineering and Life Sciences | |||||

Content | High Performance Computing: - Advanced topics in shared-memory programming - Advanced topics in MPI - GPU architectures and CUDA programming Uncertainty Quantification: - Uncertainty quantification under parametric and non-parametric modeling uncertainty - Bayesian inference with model class assessment - Markov Chain Monte Carlo simulation | |||||

Lecture notes | https://www.cse-lab.ethz.ch/teaching/hpcse-ii_fs20/ Class notes, handouts | |||||

Literature | - Class notes - Introduction to High Performance Computing for Scientists and Engineers, G. Hager and G. Wellein - CUDA by example, J. Sanders and E. Kandrot - Data Analysis: A Bayesian Tutorial, D. Sivia and J. Skilling - An introduction to Bayesian Analysis - Theory and Methods, J. Gosh, N. Delampady and S. Tapas - Bayesian Data Analysis, A. Gelman, J. Carlin, H. Stern, D. Dunson, A. Vehtari and D. Rubin - Machine Learning: A Bayesian and Optimization Perspective, S. Theodorides | |||||

Prerequisites / Notice | Students must be familiar with the content of High Performance Computing for Science and Engineering I (151-0107-20L) | |||||

101-0521-10L | Machine Learning for Predictive Maintenance Applications The number of participants in the course is limited to 25 students. Students interested in attending the lecture are requested to upload their transcript and a short motivation responding the following two questions (max. 200 words): -How does this course fit to the other courses you have attended so far? -How does the course support you in achieving your goal? The following link can be used to upload the documents. https://polybox.ethz.ch/index.php/s/3S9ZlyxQTiOS3fM | W | 8 credits | 4G | O. Fink | |

Abstract | The course aims at developing machine learning algorithms that are able to use condition monitoring data efficiently and detect occurring faults in complex industrial assets, isolate their root cause and ultimately predict the remaining useful lifetime. | |||||

Objective | Students will - be able to understand the main challenges faced by predictive maintenance systems - learn to extract relevant features from condition monitoring data -learn to select appropriate machine learning algorithms for fault detection, diagnostics and prognostics -learn to define the learning problem in way that allows its solution based on existing constrains such as lack of fault samples. - learn to design end-to-end machine learning algorithms for fault detection and diagnostics -be able to evaluate the performance of the applied algorithms. At the end of the course, the students will be able to design data-driven predictive maintenance applications for complex engineered systems from raw condition monitoring data. | |||||

Content | Early and reliable detection, isolation and prediction of faulty system conditions enables the operators to take recovery actions to prevent critical system failures and ensure a high level of availability and safety. This is particularly crucial for complex systems such as infrastructures, power plants and aircraft engines. Therefore, their system condition is increasingly tightly monitored by a large number of diverse condition monitoring sensors. With the increased availability of data on system condition on the one hand, and the increased complexity of explicit system physics-based models on the other hand, the application of data-driven approaches for predictive maintenance has been recently increasing. This course provides insights and hands-on experience in selecting, designing, optimizing and evaluating machine learning algorithms to tackle the challenges faced by maintenance systems of complex engineered systems. Specific topics include: -Introduction to condition monitoring and predictive maintenance systems -Feature extraction and selection methodology -Machine learning algorithms for fault detection and fault isolation -End-to-end learning architectures (including feature learning) for fault detection and fault isolation -Unsupervised and semi-supervised learning algorithms for predictive maintenance -Machine learning algorithms for prediction of the remaining useful life -Performance evaluation -Predictive maintenance systems at fleet level -Domain adaptation for fault diagnostics -Introduction to decision support systems for maintenance applications | |||||

Lecture notes | Slides and other materials will be available online. | |||||

Literature | Relevant scientific papers will be discussed in the course. | |||||

Prerequisites / Notice | Strong analytical skills. Programming skills in python are strongly recommended. | |||||

103-0848-00L | Industrial Metrology and Machine Vision Number of participants limited to 30. | W | 4 credits | 3G | K. Schindler, A. Wieser | |

Abstract | This course introduces contact and non-contact techniques for 3D coordinate, shape and motion determination as used for 3D inspection, dimensional control, reverse engineering, motion capture and similar industrial applications. | |||||

Objective | Understanding the physical basis of photographic sensors and imaging; familiarization with a broader view of image-based 3D geometry estimation beyond the classical photogrammetric approach; understanding the concepts of measurement traceability and uncertainty; acquiring an overview of general 3D image metrology including contact and non-contact techniques (coordinate measurement machines; optical tooling; laser-based high-precision instruments). | |||||

Content | CCD and CMOS technology; structured light and active stereo; shading models, shape from shading and photometric stereo; shape from focus; laser interferometry, laser tracker, laser radar; contact and non-contact coordinate measurement machines; optical tooling; measurement traceability, measurement uncertainty, calibration of measurement systems; 3d surface representations; case studies. | |||||

Lecture notes | Lecture slides and further literature will be made available on the course webpage. | |||||

227-0216-00L | Control Systems II | W | 6 credits | 4G | R. Smith | |

Abstract | Introduction to basic and advanced concepts of modern feedback control. | |||||

Objective | Introduction to basic and advanced concepts of modern feedback control. | |||||

Content | This course is designed as a direct continuation of the course "Regelsysteme" (Control Systems). The primary goal is to further familiarize students with various dynamic phenomena and their implications for the analysis and design of feedback controllers. Simplifying assumptions on the underlying plant that were made in the course "Regelsysteme" are relaxed, and advanced concepts and techniques that allow the treatment of typical industrial control problems are presented. Topics include control of systems with multiple inputs and outputs, control of uncertain systems (robustness issues), limits of achievable performance, and controller implementation issues. | |||||

Lecture notes | The slides of the lecture are available to download. | |||||

Literature | Skogestad, Postlethwaite: Multivariable Feedback Control - Analysis and Design. Second Edition. John Wiley, 2005. | |||||

Prerequisites / Notice | Prerequisites: Control Systems or equivalent | |||||

227-0224-00L | Stochastic Systems | W | 4 credits | 2V + 1U | F. Herzog | |

Abstract | Probability. Stochastic processes. Stochastic differential equations. Ito. Kalman filters. St Stochastic optimal control. Applications in financial engineering. | |||||

Objective | Stochastic dynamic systems. Optimal control and filtering of stochastic systems. Examples in technology and finance. | |||||

Content | - Stochastic processes - Stochastic calculus (Ito) - Stochastic differential equations - Discrete time stochastic difference equations - Stochastic processes AR, MA, ARMA, ARMAX, GARCH - Kalman filter - Stochastic optimal control - Applications in finance and engineering | |||||

Lecture notes | H. P. Geering et al., Stochastic Systems, Measurement and Control Laboratory, 2007 and handouts | |||||

227-0690-11L | Advanced Topics in Control (Spring 2020)New topics are introduced every year. | W | 4 credits | 2V + 2U | G. Banjac | |

Abstract | Advanced Topics in Control (ATIC) covers advanced research topics in control theory. It is offered each Spring semester with the topic rotating from year to year. Repetition for credit is possible, with consent of the instructor. | |||||

Objective | During Spring 2020 the course will cover a range of topics in large-scale convex optimization. The students should be able to apply various numerical methods to solve large-scale optimization problems arising in control, machine learning, signal processing, and finance. | |||||

Content | Convex analysis and methods for large-scale optimization. Topics will include: convex sets and functions ; duality theory ; optimality and infeasibility conditions ; structured optimization problems ; gradient-based methods ; operator splitting methods ; distributed and decentralized optimization ; applications in various research areas. | |||||

Lecture notes | Copies of the projection slides will be made available on the course Moodle platform. | |||||

Literature | The course will be largely based on the Large-Scale Convex Optimization course taught at Lund University: https://archive.control.lth.se/ls-convex-2015/ | |||||

Prerequisites / Notice | Sufficient mathematical maturity, in particular in linear algebra and analysis. | |||||

252-0220-00L | Introduction to Machine Learning Limited number of participants. Preference is given to students in programmes in which the course is being offered. All other students will be waitlisted. Please do not contact Prof. Krause for any questions is this regard. If necessary, please contact studiensekretariat@inf.ethz.ch | W | 8 credits | 4V + 2U + 1A | A. Krause | |

Abstract | The course introduces the foundations of learning and making predictions based on data. | |||||

Objective | The course will introduce the foundations of learning and making predictions from data. We will study basic concepts such as trading goodness of fit and model complexitiy. We will discuss important machine learning algorithms used in practice, and provide hands-on experience in a course project. | |||||

Content | - Linear regression (overfitting, cross-validation/bootstrap, model selection, regularization, [stochastic] gradient descent) - Linear classification: Logistic regression (feature selection, sparsity, multi-class) - Kernels and the kernel trick (Properties of kernels; applications to linear and logistic regression); k-nearest neighbor - Neural networks (backpropagation, regularization, convolutional neural networks) - Unsupervised learning (k-means, PCA, neural network autoencoders) - The statistical perspective (regularization as prior; loss as likelihood; learning as MAP inference) - Statistical decision theory (decision making based on statistical models and utility functions) - Discriminative vs. generative modeling (benefits and challenges in modeling joint vy. conditional distributions) - Bayes' classifiers (Naive Bayes, Gaussian Bayes; MLE) - Bayesian approaches to unsupervised learning (Gaussian mixtures, EM) | |||||

Literature | Textbook: Kevin Murphy, Machine Learning: A Probabilistic Perspective, MIT Press | |||||

Prerequisites / Notice | Designed to provide a basis for following courses: - Advanced Machine Learning - Deep Learning - Probabilistic Artificial Intelligence - Seminar "Advanced Topics in Machine Learning" | |||||

252-0526-00L | Statistical Learning Theory | W | 7 credits | 3V + 2U + 1A | J. M. Buhmann, C. Cotrini Jimenez | |

Abstract | The course covers advanced methods of statistical learning: - Variational methods and optimization. - Deterministic annealing. - Clustering for diverse types of data. - Model validation by information theory. | |||||

Objective | The course surveys recent methods of statistical learning. The fundamentals of machine learning, as presented in the courses "Introduction to Machine Learning" and "Advanced Machine Learning", are expanded from the perspective of statistical learning. | |||||

Content | - Variational methods and optimization. We consider optimization approaches for problems where the optimizer is a probability distribution. We will discuss concepts like maximum entropy, information bottleneck, and deterministic annealing. - Clustering. This is the problem of sorting data into groups without using training samples. We discuss alternative notions of "similarity" between data points and adequate optimization procedures. - Model selection and validation. This refers to the question of how complex the chosen model should be. In particular, we present an information theoretic approach for model validation. - Statistical physics models. We discuss approaches for approximately optimizing large systems, which originate in statistical physics (free energy minimization applied to spin glasses and other models). We also study sampling methods based on these models. | |||||

Lecture notes | A draft of a script will be provided. Lecture slides will be made available. | |||||

Literature | Hastie, Tibshirani, Friedman: The Elements of Statistical Learning, Springer, 2001. L. Devroye, L. Gyorfi, and G. Lugosi: A probabilistic theory of pattern recognition. Springer, New York, 1996 | |||||

Prerequisites / Notice | Knowledge of machine learning (introduction to machine learning and/or advanced machine learning) Basic knowledge of statistics. |

- Page 1 of 2 All