IfI Summer School 2018 on Machine Learning

The 2018 IfI Summer School is a week-long event for PhD students and research assistants in informatics and related fields, where invited experts teach a number of different topics in day-long courses on Machine Learning.

Dates and Location

The summer school will take place June 25-29, 2018 at the University of Zurich, Department of Informatics.

Course location:
Binzmühlestrasse 14
8050 Zurich
Class rooms BIN 2.A.01 und 2.A.10

Courses will be held from 9 a.m. - 5 p.m. (check-in starts at 8:45) with coffee and lunch breaks.

List of Courses

Day Course Instructor ECTS credits
MON, June 25 Introduction to Machine Learning Prof. Dr. Andreas Krause 0.5 Doctoral
TUE, June 26 Deep Learning with Py Torch Dr. François Fleuret 0.5 Doctoral
TUE, June 26

Deep Learning for Natural Language Processing

Dr. Felipe Bravo-Marquez 0.5 Doctoral
WED, June 27

Sensorimotor Control and Deep Reinforcement Learning

Dr. Alexey Dosovitskiy 0.5 Doctoral
WED, June 27 Learning with Knowledge Graphs  and Causal Reasoning for
Artificial Intelligence
Prof. Dr. Volker Tresp 0.5 Doctoral
THU, June 28 Tensor Methods for Large Scale Machine-Learning Jean Kossaifi 0.5 Doctoral
THU, June 28 Machine Learning in Market Design Prof. Dr. Jason Hartline 0.5 Doctoral
FRI, June 29 Machine Learning for Software Engineering Prof. Dr. Alberto Bacchelli 0.5 Doctoral
FRI, June 29

Individual and Group Fairness in Machine Learning

Prof. Dr. Aaron Roth 0.5 Doctoral

Please note: Every course covers a full day. On some weekdays, there are parallel sessions of two courses per day. Please make sure to register for one course per weekday only. 

Daily schedule (subject to change as needed by instructors)

08:45 - 09:00 Check-in
09:00 - 10:15 Instruction
10:15 - 10:45 Coffee break
10:45 - 12:00 Instruction
12:00 - 13:00 Lunch (@mensa, not included in cost)
13:00 - 15:00 Instruction
15:00 - 15:30 Coffee break
15:30 - 17:00 Instruction

All registered students are also invited to attend the summer school social event. Details to follow.


The Summer School is primarily targeted towards doctoral students in computer science and related fields from the University of Zurich as well as other universities. Registration is free for IfI research assistants, IfI doctoral students, and IFI postdocs. For all other participants, fees are 180 CHF for the entire five-day summer school, or 40 CHF for individual courses. Attendance will be capped at 40 people per course.

Please note that we cannot issue any invitation letters for visa issues.

Registration is now closed.

Payment Method

The fee will be paid on site in cash only at the check-in desk outside the classrooms every day between 8:45 and 9:00 a.m.

ECTS Credits

For UZH students, you can find the ECTS credit awarded by each course in the overview above. For non-IfI students who would like to acquire cedits, you need to talk with the person who is in charge of credit transferring at your university first and find out if the ECTS credits awarded by IfI at UZH are accepted/recognized by your university.

Courses and Instructors

June 25

Introduction to Machine Learning

Course Description

This course will provide an introduction to basic concepts from supervised and unsupervised machine learning.  We will cover basic approaches to linear and nonlinear classification, regression, clustering and dimension reduction.  We also discuss principles of model selection and validation for optimising predictive performance.  The lectures will be accompanied by interactive demos.

Instructor Prof. Dr. Andreas Krause
Short Bio Andreas Krause is a Professor of Computer Science at ETH Zurich, where he leads the Learning & Adaptive Systems Group. He also serves as Academic Co-Director of the Swiss Data Science Center. Before that he was an Assistant Professor of Computer Science at Caltech. He received his Ph.D. in Computer Science from Carnegie Mellon University (2008) and his Diplom in Computer Science and Mathematics from the Technical University of Munich, Germany (2004). He is a Microsoft Research Faculty Fellow and a Kavli Frontiers Fellow of the US National Academy of Sciences. He received an ERC Starting Investigator grant, the Deutscher Mustererkennungspreis, an NSF CAREER award, the Okawa Foundation Research Grant recognizing top young researchers in telecommunications as well as the ETH Golden Owl teaching award. His research on machine learning and adaptive systems has received awards at several premier conferences and journals. Andreas Krause is currently serving as Program Co-Chair for ICML 2018, and is regularly serving as Area Chair or Senior Program Committee member for ICML, NIPS, AAAI and IJCAI, and as Action Editor for the Journal of Machine Learning Research.

June 26

Deep Learning with Py Torch

Course Description This course will first provide a general introduction to deep learning and its relation to classical machine learning. We will then look at techniques and concepts specific to the design of large-scale computation-intensive models, and illustrate these notions with examples in the PyTorch framework.
Instructor Dr. François Fleuret
Short Bio François Fleuret got a PhD in Mathematics from INRIA and the University of Paris VI in 2000, and an Habilitation degree in Mathematics from the University of Paris XIII in 2006. He is the head of the Machine Learning group at the Idiap Research Institute, Switzerland, since 2007, and adjunct faculty at the École Polytechnique Fédérale de Lausanne (EPFL) since 2011, where he teaches machine learning. He has published more than 80 papers in peer-reviewed international conferences and journals. He is Associate Editor of the IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) since 2012, served as Area Chair for NIPS (2012, 2014, 2016, 2017) and ICCV (2012) and in the program committee of many top-tier international conferences in machine learning and computer vision. He is member of the Electrical Engineering Doctoral Program Committee at EPFL, and was or is expert for multiple funding agencies (Swiss National Science Foundation, European Research Council, Austrian Science Fund, Netherlands Organization for Scientific Research, French National Research Agency, Research Council of the Academy of Finland, US National Science Foundation). His main research interest is machine learning, with a particular focus on computational aspects and small sample learning, and applications in computer vision.

June 26

Deep Learning for Natural Language Processing

Course Description Sentiment analysis is the task of automatically classifying text into sentiment categories such as positive and negative. Until 2014, state-of-the-art solutions to this problem relied on shallow learning schemes based on hand-crafted features and linear machine learning models. Deep neural networks architectures had became very popular in the computer vision community due to its success for detecting objects (“cat”, “bicycles”) regardless of its position in the image. These approaches have also been recently adopted for many natural language processing (NLP) tasks, including sentiment analysis, with successful results. In this tutorial, we use sentiment analysis as a case study for introducing modern neural network architectures for NLP, including word embeddings, convolutional neural networks, and recurrent neural networks. No previous linguistic knowledge is required. Basic understanding of mathematical concepts such as functions, matrices, and derivatives may be helpful but is not essential.
Instructor Dr. Felipe Bravo-Marquez  
Short Bio Felipe Bravo-Marquez is a research fellow in the Machine Learning Group at the University of Waikato, New Zealand. He received his PhD degree from the University of Waikato. Previously, he received two engineering degrees in the fields of computer science and industrial engineering, and a masters degree in computer science, all from the University of Chile. He worked for three years as a research engineer at Yahoo! Labs Latin America. His main areas of interest are: data mining, natural language processing, information retrieval, and sentiment analysis. His full list of publications is available here: https://www.cs.waikato.ac.nz/~fbravoma/

June 27

Sensorimotor Control and Deep Reinforcement Learning

Course Description Sensorimotor control - producing useful actions based on sensory inputs - is a crucial capability for intelligent agents, both natural and artificial. Recently, deep reinforcement learning has shown promising results on this task, for instance, in game playing (Atari and Go) and robotics (vision-based manipulation). In this course we will review a variety of modern approaches to sensorimotor control, with a focus on deep learning and deep reinforcement learning. On the deep reinforcement learning side, we will go from fundamentals to cutting edge algorithms, show successful applications, but also highlight shortcomings and failures. We will discuss algorithms deployed both in simulation and in the real world, as well as the exciting topic of transfer from simulation to the real world.
Instructor Dr. Alexey Dosovitskiy
Short Bio Alexey Dosovitskiy is a research scientist at Intel Intelligent Systems Lab in Munich, Germany. His recent research focuses on deep learning, sensorimotor control, and simulation, with a particular focus on driving and navigation. Previously, he was a postdoctoral researcher at the Computer Vision Group of Prof. Thomas Brox at the University of Freiburg in Germany, contributing to multiple subfields at the intersection of deep learning and computer vision: unsupervised feature learning, image generation with neural networks, motion and 3D structure estimation with neural networks. He received MSc and PhD degrees in mathematics from Moscow State University in 2009 and 2012 respectively.

June 27

Learning with Knowledge Graphs  and Causal Reasoning for Artificial Intelligence

Course Description

In the first part I will cover recent work on learning with knowledge graphs. In the second part I will cover the two main causal theories of relevance for AI: Pearl's theory of causal and counterfactual

inference and Rubin's causal model. Link to detailed description (PDF, 139 KB)
Instructor Prof. Dr. Volker Tresp
Short Bio Volker Tresp received a Diploma degree from the University of Goettingen, Germany, in 1984 and the M.Sc. and Ph.D. degrees from Yale University, New Haven, CT, in 1986 and 1989 respectively. Since 1989 he is the head of various research teams in machine learning at Siemens, Research and Technology.  He filed more than 70 patent applications and was inventor of the year of Siemens in 1996. He has published more than 150 scientific articles and administered over 20 Ph.D. theses. The company Panoratio is a spin-off out of his team.  His research focus in recent years has been „Machine Learning in Information Networks“ for modelling Knowledge Graphs, medical decision processes and sensor networks. He is the coordinator of one of the first nationally funded Big Data projects for the realization of „Precision Medicine“.   Since 2011 he is also a Professor at the Ludwig Maximilian University of Munich where he teaches an annual course on Machine Learning.

June 28

Tensor Methods for Large Scale Machine-Learning

Course Description Tensors, or multi-way arrays, are a generalisation of the concept of matrix to higher dimensions and can encode higher order relationships in data. This course will provide a practical introduction to tensor methods, tensor decomposition and regression, and how to combine tensor methods with Deep Learning. We will demonstrate practical implementation of the algorithms in Python, with TensorLy and PyTorch.
Instructor Jean Kossaifi
Short Bio Jean Kossaifi's research interests are primarily focused on the areas of machine learning, computer vision and pattern recognition, with applications in human-computer interaction and facial affect estimation. He is particularly interested in tensor methods and how to combine them with deep learning. He is currently a Research Assistant at Imperial College London where he finishes his PhD this summer. He has been a visiting researcher in Oregon State University, and most recently Scientist at Amazon AI in Palo Alto, California, working on tensor methods. Jean also holds an MSc in Advanced Computing from Imperial College London, a French Engineering Diploma/MSc in applied mathematics, finance and computing, as well as a BSc in Advanced Mathematics.

June 28

Machine Learning in Market Design

Course Description

This course covers machine learning and mechanism design in the context of online markets.  In online markets; which include auctions, ride-sharing, online dating, advertising, and other applications; a market maker aims to provide a market mechanism that achieves good market outcomes even under the expectation of strategization on the part of the participants in the market.  The main themes of the course are a review of mechanism design, the role of machine learning in the design of mechanisms, and the role of machine learning in the participants’ optimization of their strategies.  Topics include: auction theory, game theoretic equilibrium, econometric inference, sample complexity, and expert and multi-armed bandit algorithms.     

Instructor Prof. Dr. Jason Hartline
Short Bio Prof. Hartline received his Ph.D. in 2003 from the University of Washington under the supervision of Anna Karlin. He was a postdoctoral fellow at Carnegie Mellon University under the supervision of Avrim Blum; and subsequently a researcher at Microsoft Research in Silicon Valley. He joined Northwestern University in 2008 where he is an associate professor of computer science. Prof. Hartline’s research introduces design and analysis methodologies from computer science to understand and improve outcomes of economic systems. Optimal behavior and outcomes in complex environments are complex and, therefore, should not be expected; instead, the theory of approximation can show that simple and natural behaviors are approximately optimal in complex environments. This approach is applied to auction theory and mechanism design in his graduate textbook "Mechanism Design and Approximation" which is under preparation.

June 29

Machine Learning for Software Engineerig

Course Description Our society runs on software systems. From finance, to health, to agriculture, we rely on software systems; it is thus of the utmost importance to be able to do software engineering delivering high quality software with great efficiency. It turns out that machine learning can greatly help with this; indeed software engineering is a very fertile ground for machine learning: Data is collected at every step of the development process and development tasks can be formulated as machine learning problems. In this course, we are going to see how machine learning can be applied to software engineering projects and analyze recent case studies to further understand the techniques and their applications.
Instructor Prof. Dr. Alberto Bacchelli
Short Bio Alberto Bacchelli is SNSF professor in Empirical Software Engineering at UZH where he leads ZEST. His research focus is on developing and applying data science to improve software engineering, with processes, practices, and tools informed by data and actors’ needs. Prior to joining University of Zurich, Alberto Bacchelli was on the faculty at TU Delft, The Netherlands. He obtained his Ph.D. degree in 2013 at the Faculty of Informatics of the University of Lugano, Switzerland, with a thesis on Mining Unstructured Software Data. He has been research intern twice (2012 and 2013) at Microsoft Research in Redmond, USA. He also worked for CINECA (the largest Italian computing centre, one of the most important worldwide) as a software engineer.

June 29

Individual and Group Fairness in Machine Learning

Course Description This course will start by outlining why simple machine learning techniques, naively applied, can lead to unfair outcomes, even when there is no baked in bias in either the data or the algorithms specification. We will then describe two approaches to thinking about fairness in machine learning: individual definitions of fairness, and statistical definitions of fairness, together with the benefits and shortfalls of each. Finally, we will study a definition that achieves some of the best properties of both individual and statistical definitions of fairness, and analyze an oracle-efficient algorithm for training the optimal classifier subject to this fairness constraint.
Instructor Prof. Dr. Aaron Roth
Short Bio Aaron Roth is the class of 1940 Bicentennial Term associate professor of Computer and Information Sciences at the University of Pennsylvania, affiliated with the Warren Center for Network and Data Science, and co-director of the Networked and Social Systems Engineering (NETS) program. Previously, he received his PhD from Carnegie Mellon University and spent a year as a postdoctoral researcher at Microsoft Research New England. He is the recipient of a Presidential Early Career Award for Scientists and Engineers (PECASE) awarded by President Obama in 2016, an Alfred P. Sloan Research Fellowship, an NSF CAREER award, and a Yahoo! ACE award.  His research focuses on the algorithmic foundations of data privacy, algorithmic fairness, game theory and mechanism design, learning theory, and the intersections of these topics.  Together with Cynthia Dwork, he is the author of the book “The Algorithmic Foundations of Differential Privacy.”