ECML/PKDD Summer School (EPSS'19)

Navigation (dev mode): Home | Speakers | Program | Location | Application | Contact

EPSS’19 is on ‘Machine Learning and Data Mining for Geo-Spatial Data/Volunteered Geographic Information, Quality of Experience and Human-Computer Interaction.’

Key facts

Date: 11. Sep. 2019 - 16. Sep. 2019
Location: University of Würzburg
Organizers: Martin Becker, Martin Fischbach, Matthias Hirth
Supported by: ECML/PKDD

Introduction

Machine Learning and Data Mining techniques have become one of the main analysis, modeling and prediction tools for a multitude of application scenarios. The corresponding methodology becomes more sophisticated and is adapted to domain specific problem settings. As part of the ‘ECML PKDD Summer School’ (EPSS19), we aim to foster and support this development.

For this, we provide extensive foundations in state-of-the-art machine learning and data mining methods covering areas like time series analysis, deep neural networks, and automated machine learning. At the same time, the summer school features tracks to delve deeper into domain-specific methods, algorithms, and applications. We focus on three specific topics: (1) geo-spatial data and volunteered geographic information, (2) quality of experience, and (3) human-computer interaction. This curriculum is accompanied by many hands-on sessions with several highly esteemed lecturers as well as the opportunity to work on innovative paper projects together with motivated Ph.D. students from around the world.

Target audience

Any Ph.D. student interested in machine learning and data mining is more than welcome, especially when working in the field of geo-spatial data, volunteered geographical information, human-computer interaction, natural interfaces, virtual-, augmented-, and mixed reality, social robots, quality assessment, subjective user studies and real-time interactive systems.

Master students, post-docs, and researchers working in the same area are welcome to participate as well (upon availability).

Concept of the Summer School

The summer school has a strong focus on practical aspects. The participants will get to know and learn to apply various machine learning and data mining techniques in their field of research. Therefore, the summer school has two parts. During the first two to three days, the curriculum covers a general introduction to machine learning and data mining. Afterwards, the summer school features three tracks with lectures and hands-on sessions aiming at the three specific topics: (1) geo-spatial data and volunteered geographic information, (2) quality of experience, and (3) human-computer interaction, divided into human behavior analysis and understanding such input of high variability for natural interactions. Through the whole summer school, the theoretical background is provided by lectures (40% of the summer school) by well-known experts in their field. In hands-on sessions, the theoretical knowledge will be put into practice (30%). All participants will work on concrete research topics in smaller groups (4-5 participants, guided by one senior tutor).

The concept of this summer school is to form research groups of Ph.D. students supervised by a tutor. Thereby, the Ph.D. students will be grouped according to their research interests; the group will then define the concrete topics during the summer school. Before the summer school, these groups will already start discussing potential research questions and how to investigate them. A poster presentation of the group’s ideas will take place during the summer school in order to foster collaboration and innovation between all groups.

  • Lectures (40%)
  • Hands-on sessions (30%)
  • Group work on research topics (30%)

During the summer school, the groups will get time to jointly research on their common topic with the support and help of the tutor. The results will be presented in a joint session with the ECML/PKDD conference on the last day and may be published in a joint publication of the research group.

List of topics (preliminary)

Main Track: Machine learning and data mining basics

  • Time series and sequence analysis
  • Automated machine learning
  • Deep neural networks
  • Additional topics will be announced soon

Application Track 1: Geo-spatial data and volunteered geographical information

  • Modeling geo-spatial variables and human behavior
  • Remote sensing
  • Volunteered geographical information for machine learning
  • Visualization

Application Track 2: Human-computer interaction

  • Intelligent user interfaces
  • Social computing
  • Contextual computing
  • Adaptive and cognitive systems
  • Mobile interactive systems
  • Human shape and motion analysis
  • Marker-less optical motion capture
  • Visual scene understanding
  • 3D image analysis and synthesis
  • Vision and language for explainable artificial intelligence

Application Track 2: Quality of Experience

  • QoE of immersive applications and games
  • QoE of image and video quality
  • QoE of haptics and tactile internet
  • Analytical and data-driven QoE models