Knowledge Management & Discovery Lab

Logo Knowledge Management & Discovery Lab

KMD stands for "Knowledge Management and Discovery" .

The KMD Lab is part of the department Technical and Business Information Systems (ITI)
.


The KMD Lab has been established in February 2003.

 

In the KMD lab, we develop and apply data mining methods for dynamic environments, with particular emphasis on:

  • Machine Learning methods for streams and time series with gaps – prediction and feature contribution
  • Parsimonious usage of data and features – cost-aware active feature acquisition methods
  • Design of human-understandable solutions


Our application areas are:

More on our research can be found here.

Our research is reflected in our teaching curriculum, which is built around the topic of data mining: Students learn underpinnings of data mining in all bachelor courses we offer. In the mandatory courses ITO and WMS of the Bachelor Wirtschaftsinformatik degree, we focus on mining for business applications. In the Recommenders course, we elaborate on the mining methods for static and stream recommenders.

In the courses Data Mining I (two variants, one for bachelor degrees, one for master degrees), students learn fundamentals on algorithms, model evaluation and data preparation. In Data Mining II, students learn learning methods for timestamped data. In the seminars, team projects and individual projects, students learn to design and apply mining and machine learning methods in realistic applications, and they get involved in our research - in team projects and individual projects. Our courses can be found under Study.


 

News

'Best paper award' at AIME 2022

17.06.2022 -

Miro Schleicher has received the Marco Ramoni best paper award at the 20th Artificial Intelligence in Medicine (AIME) conference for his paper 'When can I expect the mHealth user to return? Prediction meets time series with gaps' (Miro Schleicher, Rüdiger Pryss, Winfried Schlee and Myra Spiliopoulou).

This work is within the frame of the UNITI project that encompasses machine learning methods for choosing the best treatment for each tinnitus patient. Treatments have an mHealth component, which assists the users towards self-empowerment and daily management of their disease. However, mHealth apps demand self-discipline; some users give up or interact very irregularly. The proposed method learns from the data of each user and from the absence of data, and it predicts if and when a user will start interacting again with the app.

aime_award

more ...

Tutorial Mining and multimodal learning from complex medical data

06.03.2023 -

The proliferation of medical data and applications has increased the need for extracting useful knowledge that can be effectively used by the healthcare domain experts. The motivation of this tutorial is to address the complexity of medical data with specific focus on their temporal nature. While earlier tutorials in both AIME as well as other related venues such as KDD and ECML/PKDD have explored the application and utility of machine learning on medical data, there has yet been limited focus on the challenges emerging from the sequential and temporal nature of such data, as well as on the need for trust by the medical practitioners.

more ...

Last Modification: 16.06.2023 - Contact Person: