Course descriptions summer term 2014

Title Lecturer Location Time
Lectures
Autonomous Robotics: Action, Perception, and Cognition Schöner NB 3/57 Thu 14:15-16:00, Begin 10.04.2014
Computational Neuroscience: Vision and Memory Wiskott NB 3/57 Tue 12:15-13:45, Begin 08.04.2014
Mathematics for Modeling and Data Analysis Wiskott NB 3/57 Thu 12:15-13:45, Begin 10.04.2014
Vision in Man and Machine Würtz HZO 100 Fri 12:15-13:45, Begin 11.04.2014
Machine Learning: Supervised Methods Glasmachers NB 3/57 Mon 12:15-14:00, Begin 07.04.2014
Digital Image Processing Winter, Schlipsing, Houben NB 3/99 Tue 10:15-11:45, Begin 08.04.2014
Exercises for lectures
Autonomous Robotics: Action, Perception, and Cognition Schöner NB 3/57 Thu 16:15-17:00, Begin 17.04.2014
Vision in Man and Machine Würtz HZO 100 Wed 14:15-15:00 or 15:15-16:00, Begin 16.04.2014
Computational Neuroscience: Vision and Memory Wiskott NB 3/57 Tue 10:30-12:00, Begin 15.04.2014
Digital Image Processing Winter, Schlipsing, Houben NB 3/57 Mon 11:15-12:15 or 14:15-15:15, Begin 14.04.2014
Mathematics for Modeling and Data Analysis Wiskott NB 3/57 Thu 10:30-12:00, Begin 17.04.2014
Machine Learning: Supervised Methods Glasmachers NB 3/57 Fri 11:15-12:45, Begin 11.04.2014
Seminars
Selected Topics of Neuroinformatics: Intelligence Würtz, co-workers NB 3/72 Tue 16:15-18:00, Begin 08.04.2014, Attendance required on 08.04.2014!
Research Seminar
(For Master- and PhD-Students of the research group)
Würtz NB 3/72 Tue 14:00-16:00
Methods of Neuroinformatics Schöner, co-workers by arrangement by arrangement
Neurocomputing Colloquium Schöner, Dinse, Würtz, Wiskott, Jancke, co-workers by arrangement by arrangement
Application of Optical Imaging Methods and Population Approaches in Visual Cortical Areas Jancke by arrangement by arrangement
Dimensionality reduction and learning of manifolds Glasmachers NB 3/72 Wed 10:15-11:45, Begin 09.04.2014, Registration required by email (tobias.glasmachers@ini.rub.de)
Colloquium: Current topics in neuroinformatics Schöner, Wiskott, Würtz, Dinse, Jancke, co-workers NB 3/57 Wed 12:30-14:00
Lab exercises
Computer Vision Winter NB 02/77 1 week
Begin: 15.09.2014
Preliminary meeting: 04.09.2014, 10:15, NB 3/57, Limited number of participants! Attention, dates have changed!
Autonomous Robotics Schöner NB 02/77 1 week
Begin: 08.09.2014
Preliminary meeting: 04.09.2014, 10:15, NB 3/57, Limited number of participants! Attention, dates have changed!
Methods of Psychophysics Dinse   5 days, by arrangement
S-Block   (for Biology students)

Perceptual Learning

LV-NR 310749

Dinse, N.N. NB 3/72 6 weeks, includes lectures and seminars

Neurophysiology of Sensory Processing

LV-NR 310549

Dinse, Jancke NB 3/72 6 weeks, includes lectures and seminars

Theory and Physiology of Neuronal Networks

LV-NR 310649

Dinse, N.N. NB 3/72 6 weeks, includes lectures and seminars
Activation dynamics in sensory brain areas Jancke, Dinse NB 3/72 6 weeks, includes lectures and seminars
Instruction for independent scientific work
Instruction for independent scientific work Schöner, Dinse, Würtz, Wiskott, Jancke    

 

Lectures

Lecture Vision in Man and Machine

Würtz (2 HPW)

Certificate

  • Upon successful completion of the exercises (1 HPW)

Contents

  • This lecture treats aspects of vision from the computer science, psychophysics, and neurobiological point of view. Starting from a phenomenology of human vision, biological and psychological basics, the foundations of image processing are developed, finally leading to more advanced concepts like multi-scale and wavelet analysis, and algorithms for face and object recognition. The lecture is suited for students of computer science, biology, physics, engineering, and psychology.

Literature

  • B. Jähne: Digitale Bildverarbeitung, 4. Auflage, Springer.
  • I. Roth & V. Bruce: perception and Representation, 2nd edition, Open University Press.
  • Rainer Guski: Wahrnehmen - ein Lehrbuch, Kohlhammer.

Computational Neuroscience: Vision and Memory

Wiskott (2 HPW lecture + 2 HPW tutorial)

Certificate

  • Upon successful participation in the tutorial

Contents

  • This lecture presents models of selforganization in neural systems, in particular addressing vision (receptive fields, neural maps, invariances, attention) and associative memory (Hopfield network).

Literature

  • Mostly lecture notes will be provided.

Lecture homepage

  • Computational Neuroscience: Vision and Memory

Lecture Mathematics for Modeling and Data Analysis

Wiskott (2 HPW lecture + 2 HPW tutorial)

Certificate

  • Upon successful participation in the tutorial

Contents

  • This course covers mathematical methods that are relevant for modeling and data analysis. Particular emphasis will be put on an intuitive understanding as is required for a creative command of mathematics. The following topics will be covered: Functions, Hilbert-Spaces, matrices as transformations, systems of linear differential equations, qualitative analysis of nonlinear differential equations, Bayes theory, multiple integrals.

Literature

  • No particular recommendations. Standard text books will do.

Lecture homepage

  • Mathematics for Modeling and Data Analysis

Lecture Machine Learning: Supervised Methods

Glasmachers (2 HPW lecture + 2 HPW tutorial)

Contents

  • The field of machine learning constitutes a modern approach to artificial intelligence. It is situated in between neuroscience, statistics, robotics, and areas of application ranging all over science and engineering, medicine, economics, and many more. Machine learning algorithms automate the process of learning, thus allowing prediction and decision making machines to improve with experience.

    This lecture will cover different state-of-the-art methods in the domain of "supervised learning". Topics include classical statistical methods, neural networks, support vector machines, and nearest neighbor models. The lecture covers algorithmic as well as learning theoretical aspects.

    The 2 hours/week lecture is accompanied by a 2 hours/week practical course. It will be held either in German or in English, depending on the audience. Most of the course material will be in English.

    course homepage

Seminars

Selected Topics of Neuroinformatics: Intelligence

Würtz

Dieses interdisziplinäre Seminar beleuchtet das Phänomen "Intelligenz" von verschiedenen Standpunkten aus. Zunächst werden biologische und psychologische Grundlagen der Intelligenz bei Mensch und Tier behandelt, dann verschiedene Ansätze, intelligente Maschinen zu bauen.

Jeder Teilnehmer hält einen Vortrag von 30 Minuten. Die Themen werden in der ersten Sitzung am 08.04.2014 vergeben.

Methods of Neuroinformatics

Schöner, Coworkers (2 HPW)

Certificate

  • Upon giving an oral presentation

Contents

  • The seminar aims to enable students to study the scientific literature in the interdisciplinary research area of Neuroinformatics. Important publications and current projects will be discussed in fields as diverse as computer vision, autonomous robotics, intelligent systems, neural networks, evolutionary algorithms, and theoretical neuroscience.

Dimensionality reduction and learning of manifolds

Glasmachers

Ziel der Reduktion der Dimensionalität vektorieller Daten ist eine Einbettung von Daten in einen niedrigdimensionalen Raum unter Erhaltung wichtige Eigenschaften der ursprünglichen Daten, wie z. B. paarweiser Distanzen. Dimensionsreduktion hat eine Reihe wichtiger Anwendungen. So erfordert z. B. die Visualisierung von Daten eine Reduktion auf zwei oder maximal drei Dimensionen. Statistische Verfahren profitieren oft von einer niedrigen Datendimension. Allgemein sind niedrigdimensionale Daten leichter zu verstehen, zu verarbeiten, zu speichern und zu
transferieren.

In diesem Seminar wird eine Reihe von Methoden zur Dimensionsreduktion erarbeitet. Dieses Problem ist eng verwandt mit der Identifikation einer niedrigdimensionalen Untermannigfaltigkeit, die die Daten (in guter
Näherung) enthält. Im Rahmen des Seminars werden die Studierenden weitgehend selbstständig eine Reihe von Algorithmen, die den aktuellen Stand der Technik markieren, erarbeiten und vorstellen. Die Algorithmen sollen weiterhin implementiert und auf gegebenen Daten getestet werden.

Lab exercises

Computer Vision

Winter (one week, full time)

Contents

The goal of the lab exercise is to introduce to the basics of digital image processing considering current examples of computer vision.

First a short overview of programming with MATLAB is given. Then the handling of digital images is learned by implementing simple operations like mirroring and tiling. In the next step basic image processing methods, e.g. histogram equalization, image filtering, edge detection or recognition of objects with defined shape, are implemented.

Detailed information is available here.

Autonomous Robotics

Schöner (one week, full time)

Contents

The practical course gives an introduction to mobile robotics with a focus on dynamical systems approaches. In the exercises, the computing environment Matlab is used to control e-puck miniature mobile robots, equipped with a differential drive, combined infrared/proximity sensors and a video camera. The course covers elementary problems in robot odometry, use of sensors and motor control. It then teaches basic dynamic methods for robot navigation, in which the robot's sensors are used for obstacle avoidance and approach to a target location.

Each day begins with a brief introduction of the topic, then theoretical and programming exercises are solved in small groups (typically of two students).

The number of places is limited, and prior registration by e-mail is advised. The final assignment of places is done during the preliminary meeting. Interested students who do not have experience in Matlab should attend the Matlab introduction of the lab exercise Computer Vision (typically the week before this course).

Contact

sebastian.schneegans@ini.rub.de

Methods of psychophysics

Dinse (5 days)

Kommentar

  • In dieser Veranstaltung werden Grundlagen psychophysischer Methoden vorgestellt. Dazu werden unterschiedliche Typen von experimentellen Herangehensweisen vorgestellt sowie die wichtigsten statistischen Verfahren besprochen, um die damit gewonnenen Daten zu analysieren. In praktischen Übungen besteht die Möglichkeit, ausgewählte psychophysische Experimente aus dem Bereich taktiler Wahrnehmung durchzuführen. Der Kurs wird mit einem Refereat abgeschlossen.

Introduction to C++

Würtz, Glasmachers

Contents

  • Dieses Praktikum soll Studenten, die schon Java oder eine andere imperative Sprache beherrschen, den Einstieg in die Programmierung in C++ erleichtern. Diese Sprache wird für Studienprojekte und Abschlussarbeiten in 5 Arbeitsgruppen am Institut für Neuroinformatik verwendet und ist auch für die Übungen „Künstliche Neuronale Netze“ und „Sehen in Mensch und Maschine“ erforderlich.
  • Die Konzepte werden von Mitarbeitern vorgestellt und dann an praktischen Aufgaben eingeübt. Die Themen sind grob wie folgt:
     
    Grundlagen (C/C++): Kontrollstrukturen, Typsystem, Literalkonstanten, Operationen, implizite/explizite Casts, Funktionen, Deklaration/Definition, Präprozessor, Pointer und Arrays, interne/externe Bindung, Compiler-Linker-Konzept, Speicherverwaltung
     
    Klassen in C++:  Referenzen, const-Qualifizierer, Default-Parameter, Motivation: Kapselung, Abstraktion, Polymorphie, Sichtbarkeit, Konstruktor/Destruktor, Überladen von Funktionen, Kopierkonstruktor, Zuweisungsoperator, Vererbung, Überschreiben von Funktionen, virtuelle Funktionen, abstrakte Klassen, statische/dynamische Bindung, statische Elemente/Methoden
     
    Templates: Template-Funktionen, Template-Methoden, Template-Klassen, inline, explicit inline, Spezialisierung, Ausblick: Metaprogrammierung
     
    STL, Standard Template Library: cout, cin, string, fstream, vector, list, queue
    boost:  weitere Templates und Bibliotheken
    openMP: Techniken zur Parallelverarbeitung
    pragma: smart pointers

 Übersichtsseite mit Kursmaterialien