Search result: Catalogue data in Spring Semester 2020

Computer Science Master Information
Focus Courses
Focus Courses General Studies
Elective Focus Courses General Studies
NumberTitleTypeECTSHoursLecturers
252-0312-00LUbiquitous Computing Information W4 credits2V + 1AC. Holz, F. Mattern, S. Mayer
AbstractUnlike desktop computing, ubiquitous computing occurs anytime and everywhere, using any device, in any location, and in any format. Computers exist in different forms, from watches and phones to refrigerators or pairs of glasses.
Main topics: Smart environments, IoT, mobiles & wearables, context & location, sensing & tracking, computer vision on embedded systems, health monitoring, fabrication.
ObjectiveUnlike desktop computing, ubiquitous computing occurs anytime and everywhere, using any device, in any location, and in any format. Computers exist in different forms, from watches and phones to refrigerators or pairs of glasses.
Main topics: Smart environments, IoT, mobiles & wearables, context & location, sensing & tracking, computer vision on embedded systems, health monitoring, fabrication.
Lecture notesCopies of slides will be made available
LiteratureWill be provided in the lecture. To put you in the mood:
Mark Weiser: The Computer for the 21st Century. Scientific American, September 1991, pp. 94-104
252-0408-00LCryptographic Protocols Information W6 credits2V + 2U + 1AM. Hirt, U. Maurer
AbstractThe course presents a selection of hot research topics in cryptography. The choice of topics varies and may include provable security, interactive proofs, zero-knowledge protocols, secret sharing, secure multi-party computation, e-voting, etc.
ObjectiveIndroduction to a very active research area with many gems and paradoxical
results. Spark interest in fundamental problems.
ContentThe course presents a selection of hot research topics in cryptography. The choice of topics varies and may include provable security, interactive proofs, zero-knowledge protocols, secret sharing, secure multi-party computation, e-voting, etc.
Lecture notesthe lecture notes are in German, but they are not required as the entire
course material is documented also in other course material (in english).
Prerequisites / NoticeA basic understanding of fundamental cryptographic concepts
(as taught for example in the course Information Security or
in the course Cryptography Foundations) is useful, but not required.
252-0437-00LDistributed Algorithms Information W5 credits3V + 1AF. Mattern
AbstractModels of distributed computations, time space diagrams, virtual time, logical clocks and causality, wave algorithms, parallel and distributed graph traversal, consistent snapshots, mutual exclusion, election and symmetry breaking, distributed termination detection, garbage collection in distributed systems, monitoring distributed systems, global predicates.
ObjectiveBecome acquainted with models and algorithms for distributed systems.
ContentVerteilte Algorithmen sind Verfahren, die dadurch charakterisiert sind, dass mehrere autonome Prozesse gleichzeitig Teile eines gemeinsamen Problems in kooperativer Weise bearbeiten und der dabei erforderliche Informationsaustausch ausschliesslich über Nachrichten erfolgt. Derartige Algorithmen kommen im Rahmen verteilter Systeme zum Einsatz, bei denen kein gemeinsamer Speicher existiert und die Übertragungszeit von Nachrichten i.a. nicht vernachlässigt werden kann. Da dabei kein Prozess eine aktuelle konsistente Sicht des globalen Zustands besitzt, führt dies zu interessanten Problemen.
Im einzelnen werden u.a. folgende Themen behandelt:
Modelle verteilter Berechnungen; Raum-Zeit Diagramme; Virtuelle Zeit; Logische Uhren und Kausalität; Wellenalgorithmen; Verteilte und parallele Graphtraversierung; Berechnung konsistenter Schnappschüsse; Wechselseitiger Ausschluss; Election und Symmetriebrechung; Verteilte Terminierung; Garbage-Collection in verteilten Systemen; Beobachten verteilter Systeme; Berechnung globaler Prädikate.
Literature- F. Mattern: Verteilte Basisalgorithmen, Springer-Verlag
- G. Tel: Topics in Distributed Algorithms, Cambridge University Press
- G. Tel: Introduction to Distributed Algorithms, Cambridge University Press, 2nd edition
- A.D. Kshemkalyani, M. Singhal: Distributed Computing, Cambridge University Press
- N. Lynch: Distributed Algorithms, Morgan Kaufmann Publ
252-0526-00LStatistical Learning Theory Information W7 credits3V + 2U + 1AJ. M. Buhmann, C. Cotrini Jimenez
AbstractThe course covers advanced methods of statistical learning:

- Variational methods and optimization.
- Deterministic annealing.
- Clustering for diverse types of data.
- Model validation by information theory.
ObjectiveThe 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 notesA draft of a script will be provided. Lecture slides will be made available.
LiteratureHastie, 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 / NoticeKnowledge of machine learning (introduction to machine learning and/or advanced machine learning)
Basic knowledge of statistics.
252-0570-00LGame Programming Laboratory Information
In the Master Programme max. 10 credits can be accounted by Labs on top of the Interfocus Courses. Additional Labs will be listed on the Addendum.
W10 credits9PB. Sumner
AbstractThe goal of this course is the in-depth understanding of the technology and programming underlying computer games. Students gradually design and develop a computer game in small groups and get acquainted with the art of game programming.
ObjectiveThe goal of this new course is to acquaint students with the
technology and art of programming modern three-dimensional computer
games.
ContentThis course addresses modern three-dimensional computer game technology. During the course, small groups of students will design and develop a computer game. Focus will be put on technical aspects of game development, such as rendering, cinematography, interaction, physics, animation, and AI. In addition, we will cultivate creative thinking for advanced gameplay and visual effects.

The "laboratory" format involves a practical, hands-on approach with traditional lectures. We will meet once a week to discuss technical issues and to track progress. For development we use MonoGames, which is a collection of libraries and tools that facilitate game development. While development will take place on PCs, we will ultimately deployour games on the Xbox One console.

At the end of the course we will present our results to the public.
Lecture notesGame Design Workshop: A Playcentric Approach to Creating Innovative Games by Tracy Fullerton
Prerequisites / NoticeThe number of participants is limited.

Prerequisites include:

- Good programming skills (Java, C++, C#, etc.)

- CG experience: Students should have taken, at a minimum, Visual
Computing. Higher level courses are recommended, such as Introduction
to Computer Graphics, Surface Representations and Geometric Modeling,
and Physically-based Simulation in Computer Graphics.
252-0579-00L3D Vision Information W5 credits3G + 1AM. Pollefeys, V. Larsson
AbstractThe course covers camera models and calibration, feature tracking and matching, camera motion estimation via simultaneous localization and mapping (SLAM) and visual odometry (VO), epipolar and mult-view geometry, structure-from-motion, (multi-view) stereo, augmented reality, and image-based (re-)localization.
ObjectiveAfter attending this course, students will:
1. understand the core concepts for recovering 3D shape of objects and scenes from images and video.
2. be able to implement basic systems for vision-based robotics and simple virtual/augmented reality applications.
3. have a good overview over the current state-of-the art in 3D vision.
4. be able to critically analyze and asses current research in this area.
ContentThe goal of this course is to teach the core techniques required for robotic and augmented reality applications: How to determine the motion of a camera and how to estimate the absolute position and orientation of a camera in the real world. This course will introduce the basic concepts of 3D Vision in the form of short lectures, followed by student presentations discussing the current state-of-the-art. The main focus of this course are student projects on 3D Vision topics, with an emphasis on robotic vision and virtual and augmented reality applications.
252-0817-00LDistributed Systems Laboratory Information
In the Master Programme max. 10 credits can be accounted by Labs
on top of the Interfocus Courses. Additional Labs will be listed on the Addendum.
W10 credits9PG. Alonso, T. Hoefler, F. Mattern, A. Singla, R. Wattenhofer, C. Zhang
AbstractThis course involves the participation in a substantial development and/or evaluation project involving distributed systems technology. There are projects available in a wide range of areas: from web services to ubiquitous computing including as well wireless networks, ad-hoc networks, and distributed application on mobile phones.
ObjectiveStudents acquire practical knowledge about technologies from the area of distributed systems.
ContentThis course involves the participation in a substantial development and/or evaluation project involving distributed systems technology. There are projects available in a wide range of areas: from web services to ubiquitous computing including as well wireless networks, ad-hoc networks, and distributed application on mobile phones. The objecte of the project is for the students to gain hands-on-experience with real products and the latest technology in distributed systems. There is no lecture associated to the course.
For information of the course or projects available, please contact Prof. Mattern, Prof. Wattenhofer, Prof. Roscoe or Prof. G. Alonso.
252-1424-00LModels of ComputationW6 credits2V + 2U + 1AM. Cook
AbstractThis course surveys many different models of computation: Turing Machines, Cellular Automata, Finite State Machines, Graph Automata, Circuits, Tilings, Lambda Calculus, Fractran, Chemical Reaction Networks, Hopfield Networks, String Rewriting Systems, Tag Systems, Diophantine Equations, Register Machines, Primitive Recursive Functions, and more.
ObjectiveThe goal of this course is to become acquainted with a wide variety of models of computation, to understand how models help us to understand the modeled systems, and to be able to develop and analyze models appropriate for new systems.
ContentThis course surveys many different models of computation: Turing Machines, Cellular Automata, Finite State Machines, Graph Automata, Circuits, Tilings, Lambda Calculus, Fractran, Chemical Reaction Networks, Hopfield Networks, String Rewriting Systems, Tag Systems, Diophantine Equations, Register Machines, Primitive Recursive Functions, and more.
252-3005-00LNatural Language Understanding Information
Does not take place this semester.
Takes place in HS20.
W5 credits2V + 1U + 1Ato be announced
AbstractThis course presents topics in natural language processing with an emphasis on modern techniques, primarily focusing on statistical and deep learning approaches. The course provides an overview of the primary areas of research in language processing as well as a detailed exploration of the models and techniques used both in research and in commercial natural language systems.
ObjectiveThe objective of the course is to learn the basic concepts in the statistical processing of natural languages. The course will be project-oriented so that the students can also gain hands-on experience with state-of-the-art tools and techniques.
ContentThis course presents an introduction to general topics and techniques used in natural language processing today, primarily focusing on statistical approaches. The course provides an overview of the primary areas of research in language processing as well as a detailed exploration of the models and techniques used both in research and in commercial natural language systems.
LiteratureLectures will make use of textbooks such as the one by Jurafsky and Martin where appropriate, but will also make use of original research and survey papers.
252-5706-00LMathematical Foundations of Computer Graphics and Vision Information W5 credits2V + 1U + 1AM. R. Oswald, C. Öztireli
AbstractThis course presents the fundamental mathematical tools and concepts used in computer graphics and vision. Each theoretical topic is introduced in the context of practical vision or graphic problems, showcasing its importance in real-world applications.
ObjectiveThe main goal is to equip the students with the key mathematical tools necessary to understand state-of-the-art algorithms in vision and graphics. In addition to the theoretical part, the students will learn how to use these mathematical tools to solve a wide range of practical problems in visual computing. After successfully completing this course, the students will be able to apply these mathematical concepts and tools to practical industrial and academic projects in visual computing.
ContentThe theory behind various mathematical concepts and tools will be introduced, and their practical utility will be showcased in diverse applications in computer graphics and vision. The course will cover topics in sampling, reconstruction, approximation, optimization, robust fitting, differentiation, quadrature and spectral methods. Applications will include 3D surface reconstruction, camera pose estimation, image editing, data projection, character animation, structure-aware geometry processing, and rendering.
261-5120-00LMachine Learning for Health Care Information Restricted registration - show details
Number of participants limited to 150.
W5 credits3P + 1AG. Rätsch, J. Vogt, V. Boeva
AbstractThe course will review the most relevant methods and applications of Machine Learning in Biomedicine, discuss the main challenges they present and their current technical problems.
ObjectiveDuring the last years, we have observed a rapid growth in the field of Machine Learning (ML), mainly due to improvements in ML algorithms, the increase of data availability and a reduction in computing costs. This growth is having a profound impact in biomedical applications, where the great variety of tasks and data types enables us to get benefit of ML algorithms in many different ways. In this course we will review the most relevant methods and applications of ML in biomedicine, discuss the main challenges they present and their current technical solutions.
ContentThe course will consist of four topic clusters that will cover the most relevant applications of ML in Biomedicine:
1) Structured time series: Temporal time series of structured data often appear in biomedical datasets, presenting challenges as containing variables with different periodicities, being conditioned by static data, etc.
2) Medical notes: Vast amount of medical observations are stored in the form of free text, we will analyze stategies for extracting knowledge from them.
3) Medical images: Images are a fundamental piece of information in many medical disciplines. We will study how to train ML algorithms with them.
4) Genomics data: ML in genomics is still an emerging subfield, but given that genomics data are arguably the most extensive and complex datasets that can be found in biomedicine, it is expected that many relevant ML applications will arise in the near future. We will review and discuss current applications and challenges.
Prerequisites / NoticeData Structures & Algorithms, Introduction to Machine Learning, Statistics/Probability, Programming in Python, Unix Command Line

Relation to Course 261-5100-00 Computational Biomedicine: This course is a continuation of the previous course with new topics related to medical data and machine learning. The format of Computational Biomedicine II will also be different. It is helpful but not essential to attend Computational Biomedicine before attending Computational Biomedicine II.
263-3501-00LFuture Internet Information W6 credits1V + 1U + 3AA. Singla
AbstractThis course will discuss recent advances in networking, with a focus on the Internet, with topics ranging from the algorithmic design of applications like video streaming to the likely near-future of satellite-based networking.
ObjectiveThe goals of the course are to build on basic undergraduate-level networking, and provide an understanding of the tradeoffs and existing technology in the design of large, complex networked systems, together with concrete experience of the challenges through a series of lab exercises.
ContentThe focus of the course is on principles, architectures, protocols, and applications used in modern networked systems. Example topics include:

- How video streaming services like Netflix work, and research on improving their performance.
- How Web browsing could be made faster
- How the Internet's protocols are improving
- Exciting developments in satellite-based networking (ala SpaceX)
- The role of data centers in powering Internet services

A series of programming assignments will form a substantial part of the course grade.
Lecture notesLecture slides will be made available at the course Web site: Link
LiteratureNo textbook is required, but there will be regularly assigned readings from research literature, liked to the course Web site: Link.
Prerequisites / NoticeAn undergraduate class covering the basics of networking, such as Internet routing and TCP. At ETH, Computer Networks (252-0064-00L) and Communication Networks (227-0120-00L) suffice. Similar courses from other universities are acceptable too.
263-3710-00LMachine Perception Information Restricted registration - show details
Number of participants limited to 200.
W5 credits2V + 1U + 1AO. Hilliges
AbstractRecent developments in neural networks (aka “deep learning”) have drastically advanced the performance of machine perception systems in a variety of areas including computer vision, robotics, and intelligent UIs. This course is a deep dive into deep learning algorithms and architectures with applications to a variety of perceptual tasks.
ObjectiveStudents will learn about fundamental aspects of modern deep learning approaches for perception. Students will learn to implement, train and debug their own neural networks and gain a detailed understanding of cutting-edge research in learning-based computer vision, robotics and HCI. The final project assignment will involve training a complex neural network architecture and applying it on a real-world dataset of human activity.

The core competency acquired through this course is a solid foundation in deep-learning algorithms to process and interpret human input into computing systems. In particular, students should be able to develop systems that deal with the problem of recognizing people in images, detecting and describing body parts, inferring their spatial configuration, performing action/gesture recognition from still images or image sequences, also considering multi-modal data, among others.
ContentWe will focus on teaching: how to set up the problem of machine perception, the learning algorithms, network architectures and advanced deep learning concepts in particular probabilistic deep learning models

The course covers the following main areas:
I) Foundations of deep-learning.
II) Probabilistic deep-learning for generative modelling of data (latent variable models, generative adversarial networks and auto-regressive models).
III) Deep learning in computer vision, human-computer interaction and robotics.

Specific topics include: 
I) Deep learning basics:
a) Neural Networks and training (i.e., backpropagation)
b) Feedforward Networks
c) Timeseries modelling (RNN, GRU, LSTM)
d) Convolutional Neural Networks for classification
II) Probabilistic Deep Learning:
a) Latent variable models (VAEs)
b) Generative adversarial networks (GANs)
c) Autoregressive models (PixelCNN, PixelRNN, TCNs)
III) Deep Learning techniques for machine perception:
a) Fully Convolutional architectures for dense per-pixel tasks (i.e., instance segmentation)
b) Pose estimation and other tasks involving human activity
c) Deep reinforcement learning
IV) Case studies from research in computer vision, HCI, robotics and signal processing
LiteratureDeep Learning
Book by Ian Goodfellow and Yoshua Bengio
Prerequisites / NoticeThis is an advanced grad-level course that requires a background in machine learning. Students are expected to have a solid mathematical foundation, in particular in linear algebra, multivariate calculus, and probability. The course will focus on state-of-the-art research in deep-learning and will not repeat basics of machine learning

Please take note of the following conditions:
1) The number of participants is limited to 200 students (MSc and PhDs).
2) Students must have taken the exam in Machine Learning (252-0535-00) or have acquired equivalent knowledge
3) All practical exercises will require basic knowledge of Python and will use libraries such as TensorFlow, scikit-learn and scikit-image. We will provide introductions to TensorFlow and other libraries that are needed but will not provide introductions to basic programming or Python.

The following courses are strongly recommended as prerequisite:
* "Visual Computing" or "Computer Vision"

The course will be assessed by a final written examination in English. No course materials or electronic devices can be used during the examination. Note that the examination will be based on the contents of the lectures, the associated reading materials and the exercises.
263-4507-00LAdvances in Distributed Graph Algorithms
Does not take place this semester.
W6 credits3V + 1U + 1AM. Ghaffari
AbstractHow can a network of computers solve the graph problems needed for running that network?
ObjectiveThis course will familiarize the students with the algorithmic tools and techniques in local distributed graph algorithms, and overview the recent highlights in the field. This will also prepare the students for independent research at the frontier of this area.

This is a special‐topics course in algorithm design. It should be accessible to any student with sufficient theoretical/algorithmic background. In particular, it assumes no familiarity with distributed computing. We only expect that the students are comfortable with the basics of algorithm design and analysis, as well as probability theory. It is possible to take this course simultaneously with the course “Principles of Distributed Computing”. If you are not sure whether you are ready for this class or not, please consult the instructor.
ContentHow can a network of computers solve the graph problems needed for running that network?

Answering this and similar questions is the underlying motivation of the area of Distributed Graph Algorithms. The area focuses on the foundational algorithmic aspects in these questions and provides methods for various distributed systems --- e.g., the Internet, a wireless network, a multi-processor computer, etc --- to solve computational problems that can be abstracted as graph problems. For instance, think about shortest path computation in routing, or about coloring and independent set computation in contention resolution.

Over the past decade, we have witnessed a renaissance in the area of Distributed Graph Algorithms, with tremendous progress in many directions and solutions for a number of decades-old central problems. This course overviews the highlights of these results. The course will mainly focus on one half of the field, which revolves around locality and local problems. The other half, which relates to the issue of congestion and dealing with limited bandwidth in global problems, will not be addressed in this offering of the course.

The course will cover a sampling of the recent developments (and open questions) at the frontier of research of distributed graph algorithms. The material will be based on a compilation of recent papers in this area, which will be provided throughout the semester. The tentative list of topics includes:
- The shattering technique for local graph problems and its necessity
- Lovasz Local Lemma algorithms, their distributed variants, and distributed applications
- Distributed Derandomization
- Distributed Lower bounds
- Graph Coloring
- Complexity Hierarchy and Gaps
- Primal-Dual Techniques
Prerequisites / NoticeThe class assumes no knowledge in distributed algorithms/computing. Our only prerequisite is the undergraduate class Algorithms, Probability, and Computing (APC) or any other course that can be seen as the equivalent. In particular, much of what we will discuss uses randomized algorithms and therefore, we will assume that the students are familiar with the tools and techniques in randomized algorithms and analysis (to the extent covered in the APC class).
263-4600-00LFormal Methods for Information Security Information W5 credits2V + 1U + 1AR. Sasse, C. Sprenger
AbstractThe course focuses on formal methods for the modelling and analysis of security protocols for critical systems, ranging from authentication protocols for network security to electronic voting protocols and online banking.
ObjectiveThe students will learn the key ideas and theoretical foundations of formal modelling and analysis of security protocols. The students will complement their theoretical knowledge by solving practical exercises, completing a small project, and using state-of-the-art tools.
ContentThe course treats formal methods mainly for the modelling and analysis of security protocols. Cryptographic protocols (such as SSL/TLS, SSH, Kerberos, SAML single-sign on, and IPSec) form the basis for secure communication and business processes. Numerous attacks on published protocols show that the design of cryptographic protocols is extremely error-prone. A rigorous analysis of these protocols is therefore indispensable, and manual analysis is insufficient. The lectures cover the theoretical basis for the (tool-supported) formal modeling and analysis of such protocols. Specifically, we discuss their operational semantics, the formalization of security properties, and techniques and algorithms for their verification.

In addition to the classical security properties for confidentiality and authentication, we will study strong secrecy and privacy properties. We will discuss electronic voting protocols, and RFID protocols (a staple of the Internet of Things), where these properties are central. The accompanying tutorials provide an opportunity to apply the theory and tools to concrete protocols. Moreover, we will discuss methods to abstract and refine security protocols and the link between symbolic protocol models and cryptographic models.

Furthermore, we will also present a security notion for general systems based on non-interference as well as language-based information flow security where non-interference is enforced via a type system.
263-4400-00LAdvanced Graph Algorithms and Optimization Information Restricted registration - show details
Number of participants limited to 30.
W5 credits3G + 1AR. Kyng
AbstractThis course will cover a number of advanced topics in optimization and graph algorithms.
ObjectiveThe course will take students on a deep dive into modern approaches to
graph algorithms using convex optimization techniques.

By studying convex optimization through the lens of graph algorithms,
students should develop a deeper understanding of fundamental
phenomena in optimization.

The course will cover some traditional discrete approaches to various graph
problems, especially flow problems, and then contrast these approaches
with modern, asymptotically faster methods based on combining convex
optimization with spectral and combinatorial graph theory.
ContentStudents should leave the course understanding key
concepts in optimization such as first and second-order optimization,
convex duality, multiplicative weights and dual-based methods,
acceleration, preconditioning, and non-Euclidean optimization.

Students will also be familiarized with central techniques in the
development of graph algorithms in the past 15 years, including graph
decomposition techniques, sparsification, oblivious routing, and
spectral and combinatorial preconditioning.
Prerequisites / NoticeThis course is targeted toward masters and doctoral students with an
interest in theoretical computer science.

Students should be comfortable with design and analysis of algorithms, probability, and linear algebra.

Having passed the course Algorithms, Probability, and Computing (APC) is highly recommended, but not formally required. If you are not
sure whether you're ready for this class or not, please consult the
instructor.
263-4656-00LDigital Signatures Information W4 credits2V + 1AD. Hofheinz
AbstractDigital signatures as one central cryptographic building block. Different security goals and security definitions for digital signatures, followed by a variety of popular and fundamental signature schemes with their security analyses.
ObjectiveThe student knows a variety of techniques to construct and analyze the security of digital signature schemes. This includes modularity as a central tool of constructing secure schemes, and reductions as a central tool to proving the security of schemes.
ContentWe will start with several definitions of security for signature schemes, and investigate the relations among them. We will proceed to generic (but inefficient) constructions of secure signatures, and then move on to a number of efficient schemes based on concrete computational hardness assumptions. On the way, we will get to know paradigms such as hash-then-sign, one-time signatures, and chameleon hashing as central tools to construct secure signatures.
LiteratureJonathan Katz, "Digital Signatures."
Prerequisites / NoticeIdeally, students will have taken the D-INFK Bachelors course "Information Security" or an equivalent course at Bachelors level.
263-5300-00LGuarantees for Machine Learning Information Restricted registration - show details W5 credits2V + 2AF. Yang
AbstractThis course teaches classical and recent methods in statistics and optimization commonly used to prove theoretical guarantees for machine learning algorithms. The knowledge is then applied in project work that focuses on understanding phenomena in modern machine learning.
ObjectiveThis course is aimed at advanced master and doctorate students who want to understand and/or conduct independent research on theory for modern machine learning. For this purpose, students will learn common mathematical techniques from statistical learning theory. In independent project work, they then apply their knowledge and go through the process of critically questioning recently published work, finding relevant research questions and learning how to effectively present research ideas to a professional audience.
ContentThis course teaches some classical and recent methods in statistical learning theory aimed at proving theoretical guarantees for machine learning algorithms, including topics in

- concentration bounds, uniform convergence
- high-dimensional statistics (e.g. Lasso)
- prediction error bounds for non-parametric statistics (e.g. in kernel spaces)
- minimax lower bounds
- regularization via optimization

The project work focuses on active theoretical ML research that aims to understand modern phenomena in machine learning, including but not limited to

- how overparameterization could help generalization ( interpolating models, linearized NN )
- how overparameterization could help optimization ( non-convex optimization, loss landscape )
- complexity measures and approximation theoretic properties of randomly initialized and
trained NN
- generalization of robust learning ( adversarial robustness, standard and robust error tradeoff )
- prediction with calibrated confidence ( conformal prediction, calibration )
Prerequisites / NoticeIt’s absolutely necessary for students to have a strong mathematical background (basic real analysis, probability theory, linear algebra) and good knowledge of core concepts in machine learning taught in courses such as “Introduction to Machine Learning”, “Regression”/ “Statistical Modelling”. It's also helpful to have heard an optimization course or approximation theoretic course. In addition to these prerequisites, this class requires a certain degree of mathematical maturity—including abstract thinking and the ability to understand and write proofs.
263-5806-00LComputational Models of Motion for Character Animation and Robotics Information W6 credits2V + 2U + 1AS. Coros, M. Bächer, B. Thomaszewski
AbstractThis course covers fundamentals of physics-based modelling and numerical optimization from the perspective of character animation and robotics applications. The methods discussed in class derive their theoretical underpinnings from applied mathematics, control theory and computational mechanics, and they will be richly illustrated using examples ranging from locomotion controllers and crowd simula
ObjectiveStudents will learn how to represent, model and algorithmically control the behavior of animated characters and real-life robots. The lectures are accompanied by programming assignments (written in C++) and a capstone project.
ContentOptimal control and trajectory optimization; multibody systems; kinematics; forward and inverse dynamics; constrained and unconstrained numerical optimization; mass-spring models for crowd simulation; FEM; compliant systems; sim-to-real; robotic manipulation of elastically-deforming objects.
Prerequisites / NoticeExperience with C++ programming, numerical linear algebra and multivariate calculus. Some background in physics-based modeling, kinematics and dynamics is helpful, but not necessary.
272-0300-00LAlgorithmics for Hard Problems Information
Does not take place this semester.
This course d o e s n o t include the Mentored Work Specialised Courses with an Educational Focus in Computer Science A.
W5 credits2V + 1U + 1A
AbstractThis course unit looks into algorithmic approaches to the solving of hard problems, particularly with moderately exponential-time algorithms and parameterized algorithms.

The seminar is accompanied by a comprehensive reflection upon the significance of the approaches presented for computer science tuition at high schools.
ObjectiveTo systematically acquire an overview of the methods for solving hard problems. To get deeper knowledge of exact and parameterized algorithms.
ContentFirst, the concept of hardness of computation is introduced (repeated for the computer science students). Then some methods for solving hard problems are treated in a systematic way. For each algorithm design method, it is discussed what guarantees it can give and how we pay for the improved efficiency. A special focus lies on moderately exponential-time algorithms and parameterized algorithms.
Lecture notesUnterlagen und Folien werden zur Verfügung gestellt.
LiteratureJ. Hromkovic: Algorithmics for Hard Problems, Springer 2004.

R. Niedermeier: Invitation to Fixed-Parameter Algorithms, 2006.

M. Cygan et al.: Parameterized Algorithms, 2015.

F. Fomin, D. Kratsch: Exact Exponential Algorithms, 2010.
272-0302-00LApproximation and Online Algorithms Information W5 credits2V + 1U + 1AH.‑J. Böckenhauer, D. Komm
AbstractThis lecture deals with approximative algorithms for hard optimization problems and algorithmic approaches for solving online problems as well as the limits of these approaches.
ObjectiveGet a systematic overview of different methods for designing approximative algorithms for hard optimization problems and online problems. Get to know methods for showing the limitations of these approaches.
ContentApproximation algorithms are one of the most succesful techniques to attack hard optimization problems. Here, we study the so-called approximation ratio, i.e., the ratio of the cost of the computed approximating solution and an optimal one (which is not computable efficiently).
For an online problem, the whole instance is not known in advance, but it arrives pieceweise and for every such piece a corresponding part of the definite output must be given. The quality of an algorithm for such an online problem is measured by the competitive ratio, i.e., the ratio of the cost of the computed solution and the cost of an optimal solution that could be given if the whole input was known in advance.

The contents of this lecture are
- the classification of optimization problems by the reachable approximation ratio,
- systematic methods to design approximation algorithms (e.g., greedy strategies, dynamic programming, linear programming relaxation),
- methods to show non-approximability,
- classic online problem like paging or scheduling problems and corresponding algorithms,
- randomized online algorithms,
- the design and analysis principles for online algorithms, and
- limits of the competitive ratio and the advice complexity as a way to do a deeper analysis of the complexity of online problems.
LiteratureThe lecture is based on the following books:

J. Hromkovic: Algorithmics for Hard Problems, Springer, 2004

D. Komm: An Introduction to Online Computation: Determinism, Randomization, Advice, Springer, 2016

Additional literature:

A. Borodin, R. El-Yaniv: Online Computation and Competitive Analysis, Cambridge University Press, 1998
401-3052-05LGraph Theory Information W5 credits2V + 1UB. Sudakov
AbstractBasic notions, trees, spanning trees, Caley's formula, vertex and edge connectivity, 2-connectivity, Mader's theorem, Menger's theorem, Eulerian graphs, Hamilton cycles, Dirac's theorem, matchings, theorems of Hall, König and Tutte, planar graphs, Euler's formula, basic non-planar graphs, graph colorings, greedy colorings, Brooks' theorem, 5-colorings of planar graphs
ObjectiveThe students will get an overview over the most fundamental questions concerning graph theory. We expect them to understand the proof techniques and to use them autonomously on related problems.
Lecture notesLecture will be only at the blackboard.
LiteratureWest, D.: "Introduction to Graph Theory"
Diestel, R.: "Graph Theory"

Further literature links will be provided in the lecture.
Prerequisites / NoticeStudents are expected to have a mathematical background and should be able to write rigorous proofs.


NOTICE: This course unit was previously offered as 252-1408-00L Graphs and Algorithms.
401-3903-11LGeometric Integer ProgrammingW6 credits2V + 1UJ. Paat
AbstractInteger programming is the task of minimizing a linear function over all the integer points in a polyhedron. This lecture introduces the key concepts of an algorithmic theory for solving such problems.
ObjectiveThe purpose of the lecture is to provide a geometric treatment of the theory of integer optimization.
ContentKey topics are:

- Lattice theory and the polynomial time solvability of integer optimization problems in fixed dimension.

- Structural properties of integer sets that reveal other parameters affecting the complexity of integer problems

- Duality theory for integer optimization problems from the vantage point of lattice free sets.
Lecture notesnot available, blackboard presentation
LiteratureLecture notes will be provided.

Other helpful materials include

Bertsimas, Weismantel: Optimization over Integers, 2005

and

Schrijver: Theory of linear and integer programming, 1986.
Prerequisites / Notice"Mathematical Optimization" (401-3901-00L)
227-0560-00LDeep Learning for Autonomous Driving Information Restricted registration - show details
Registration in this class requires the permission of the instructors. Class size will be limited to 80 students.
Preference is given to EEIT, INF and RSC students.
W6 credits3V + 2PD. Dai, A. Liniger
AbstractAutonomous driving has moved from the realm of science fiction to a very real possibility during the past twenty years, largely due to rapid developments of deep learning approaches, automotive sensors, and microprocessor capacity. This course covers the core techniques required for building a self-driving car, especially the practical use of deep learning through this theme.
ObjectiveStudents will learn about the fundamental aspects of a self-driving car. They will also learn to use modern automotive sensors and HD navigational maps, and to implement, train and debug their own deep neural networks in order to gain a deep understanding of cutting-edge research in autonomous driving tasks, including perception, localization and control.

After attending this course, students will:
1) understand the core technologies of building a self-driving car;
2) have a good overview over the current state of the art in self-driving cars;
3) be able to critically analyze and evaluate current research in this area;
4) be able to implement basic systems for multiple autonomous driving tasks.
ContentWe will focus on teaching the following topics centered on autonomous driving: deep learning, automotive sensors, multimodal driving datasets, road scene perception, ego-vehicle localization, path planning, and control.

The course covers the following main areas:

I) Foundation
a) Fundamentals of a self-driving car
b) Fundamentals of deep-learning


II) Perception
a) Semantic segmentation and lane detection
b) Depth estimation with images and sparse LiDAR data
c) 3D object detection with images and LiDAR data
d) Object tracking and motion prediction

III) Localization
a) GPS-based and Vision-based Localization
b) Visual Odometry and Lidar Odometry

IV) Path Planning and Control
a) Path planning for autonomous driving
b) Motion planning and vehicle control
c) Imitation learning and reinforcement learning for self driving cars

The exercise projects will involve training complex neural networks and applying them on real-world, multimodal driving datasets. In particular, students should be able to develop systems that deal with the following problems:
- Sensor calibration and synchronization to obtain multimodal driving data;
- Semantic segmentation and depth estimation with deep neural networks ;
- Learning to drive with images and map data directly (a.k.a. end-to-end driving)
Lecture notesThe lecture slides will be provided as a PDF.
Prerequisites / NoticeThis is an advanced grad-level course. Students must have taken courses on machine learning and computer vision or have acquired equivalent knowledge. Students are expected to have a solid mathematical foundation, in particular in linear algebra, multivariate calculus, and probability. All practical exercises will require basic knowledge of Python and will use libraries such as PyTorch, scikit-learn and scikit-image.
227-1034-00LComputational Vision (University of Zurich)
No enrolment to this course at ETH Zurich. Book the corresponding module directly at UZH.
UZH Module Code: INI402

Mind the enrolment deadlines at UZH:
Link
W6 credits2V + 1UD. Kiper
AbstractThis course focuses on neural computations that underlie visual perception. We study how visual signals are processed in the retina, LGN and visual cortex. We study the morpholgy and functional architecture of cortical circuits responsible for pattern, motion, color, and three-dimensional vision.
ObjectiveThis course considers the operation of circuits in the process of neural computations. The evolution of neural systems will be considered to demonstrate how neural structures and mechanisms are optimised for energy capture, transduction, transmission and representation of information. Canonical brain circuits will be described as models for the analysis of sensory information. The concept of receptive fields will be introduced and their role in coding spatial and temporal information will be considered. The constraints of the bandwidth of neural channels and the mechanisms of normalization by neural circuits will be discussed.
The visual system will form the basis of case studies in the computation of form, depth, and motion. The role of multiple channels and collective computations for object recognition will
be considered. Coordinate transformations of space and time by cortical and subcortical mechanisms will be analysed. The means by which sensory and motor systems are integrated to allow for adaptive behaviour will be considered.
ContentThis course considers the operation of circuits in the process of neural computations. The evolution of neural systems will be considered to demonstrate how neural structures and mechanisms are optimised for energy capture, transduction, transmission and representation of information. Canonical brain circuits will be described as models for the analysis of sensory information. The concept of receptive fields will be introduced and their role in coding spatial and temporal information will be considered. The constraints of the bandwidth of neural channels and the mechanisms of normalization by neural circuits will be discussed.
The visual system will form the basis of case studies in the computation of form, depth, and motion. The role of multiple channels and collective computations for object recognition will
be considered. Coordinate transformations of space and time by cortical and subcortical mechanisms will be analysed. The means by which sensory and motor systems are integrated to allow for adaptive behaviour will be considered.
LiteratureBooks: (recommended references, not required)
1. An Introduction to Natural Computation, D. Ballard (Bradford Books, MIT Press) 1997.
2. The Handbook of Brain Theorie and Neural Networks, M. Arbib (editor), (MIT Press) 1995.