Priv.-Doz. Dr.-Ing. habil. Georg Krempl
Knowledge Management and Discovery (KMD) Group
Priv.-Doz. Dr.-Ing. habil. Georg Krempl
Knowledge Management and Discovery (KMD) Group
Active Subtopic Detection in Multitopic Data. In Georg Krempl, Vincent Lemaire, Edwin Lughofer, and Daniel Kottke (Eds.), Proc. of the IKNOW-Workshop on Active Learning: Applications, Foundations and Emerging Trends, CEUR Workshop Proceedings, 2016. URL
Prediction-Induced Drift as New Form of Drift. In Iris Pigeot, and Eyke Hüllermeier (Eds.), Joint Statistical Meeting DAGStat2016, Big Data and Data Science Track, 2016.
Multi-Class Probabilistic Active Learning. In Gal A. Kaminka, Maria Fox, Paolo Bouquet, Eyke Hüllermeier, Virginia Dignum, Frank Dignum, and Frank van Harmelen (Eds.), ECAI, (285):586-594, IOS Press, 2016. URL
Probabilistic Active Learning for Active Class Selection. In Kory Mathewson, Kaushik Subramanian, and Robert Loftin (Eds.), Proc. of the NIPS Workshop on the Future of Interactive Learning Machines, 2016.
Learning from monitoring unlabelled data under large verification latency. In Iris Pigeot, and Eyke Hüllermeier (Eds.), Joint Statistical Meeting DAGStat2016, Big Data and Data Science Track, 2016.
Workshop on Active Learning: Applications, Foundations and Emerging Trends. In Georg Krempl, Vincent Lemaire, Edwin Lughofer, and Daniel Kottke (Eds.), CEUR Workshop Proceedings, (1707)Aachen University, Aachen, 2016. URL
Profilierung interdisziplinärer Nachwuchswissenschaftler -- Profilstudium und Summerschool Lernende Systeme / Biocomputing. In Barbara Paech (Eds.), Best practices in teaching, Deutscher Fakultätentag Informatik, 2016. URL
Inferring Delayed Neural Network Connections. In John Aston, Claudia Kirch, and Hernando Ombao (Eds.), Conference on Novel Statistical Methods in Neuroscience (NeuroStat 2016), 2016.
Investigating Exploratory Capabilities of Uncertainty Sampling using SVMs in Active Learning. In Georg Krempl, Vincent Lemaire, Edwin Lughofer, and Daniel Kottke (Eds.), Active Learning: Applications, Foundations and Emerging Trends @iKnow 2016, 25-34, 2016. URL
How to Select Information That Matters: A Comparative Study on Active Learning Strategies for Classification. Proc. of the 15th Int. Conf. on Knowledge Technologies and Data-Driven Business (i-KNOW 2015), ACM, 2015. URL
Predicting and Monitoring Changes in Scoring Data. In Jonathan Crook, David Edelman, David Hand, and Christophe Mues (Eds.), Credit Scoring and Credit Control XIV (CSCC XIV), XIVThe University of Edinburgh, 2015. URL
Probabilistic Active Learning in Datastreams. In Elisa Fromont, Tijl De Bie, and Matthijs van Leeuwen (Eds.), Advances in Intelligent Data Analysis XIV, (9385):145-157, Springer International Publishing, 2015. URL
Temporal Density Extrapolation. In Ahlame Douzal-Chouakria, José A. Vilar, Pierre-Francois Marteau, Ann Maharaj, Andrés M. Alonso, Edoardo Otranto, and Maria-Irina Nicolae (Eds.), Proc. of the 1st Int. Workshop on Advanced Analytics and Learning on Temporal Data (AALTD) co-located with ECML PKDD 2015, (1425)CEUR Workshop Proceedings, 2015. URL
When Learning Indeed Changes the World: Diagnosing Prediction-Induced Drift. In Tijl De Bie, Elisa Fromont, and Matthijs van Leeuwen (Eds.), Advances in Intelligent Data Analysis XIV - 14th Int. Symposium, IDA 2015, St. Etienne, France, (9385):XXII--XXIII, Springer, 2015.
Optimised probabilistic active learning (OPAL) For Fast, Non-Myopic, Cost-Sensitive Active Classification. In João Gama, Indrė Žliobaitė, Alípio M. Jorge, and Concha Bielza (Eds.), Machine Learning, 1-28, Springer US, 2015. URL
Clustering-Based Optimised Probabilistic Active Learning (COPAL). In Nathalie Japkowicz, and Stan Matwin (Eds.), Proc. of the 18th Int. Conf. on Discovery Science (DS 2015), (9356):101--115, Springer, 2015. URL
Predicting the post-treatment recovery of patients suffering from traumatic brain injury (TBI). Brain Informatics, 1-12, Springer Berlin Heidelberg, 2015. URL
Tagungsband der Magdeburger-Informatik-Tage, 3. Doktorandentagung 2014 (MIT 2014). In Christian Hansen, Stefan Knoll, Veit Köppen, Georg Krempl, Claudia Krull, and Eike Schallehn (Eds.), Magdeburg University, 2014. URL
Probabilistic Active Learning: A Short Proposition. In Torsten Schaub, Gerhard Friedrich, and Barry O'Sullivan (Eds.), Proceedings of the 21st European Conference on Artificial Intelligence (ECAI2014), August 18 -- 22, 2014, Prague, Czech Republic, (263)IOS Press, 2014. URL
Probabilistic Active Learning: Towards Combining Versatility, Optimality and Efficiency. In Saso Dzeroski, Pance Panov, Dragi Kocev, and Ljupco Todorovski (Eds.), Proceedings of the 17th Int. Conf. on Discovery Science (DS), Bled, Springer, 2014. URL
Are Some Brain Injury Patients Improving More Than Others?. The 2014 International Conference on Brain Informatics and Health (BIH \'14), Warsaw, Poland., 2014.
Tagungsband der Magdeburger-Informatik-Tage, 2. Doktorandentagung 2013 (MIT 2013). In Robert Buchholz, Georg Krempl, Claudia Krull, Eike Schallehn, Sebastian Stober, Frank Ortmeier, and Sebastian Zug (Eds.), Magdeburg University, 2013. URL
Drift mining in data: A framework for addressing drift in classification. Computational Statistics and Data Analysis, (57)1:377-391, 2013.
Real-World Challenges for Data Stream Mining - proceedings of the 1st International Workshop on Real-World Challenges for Data Stream Mining, RealStream 2013, Prague, Czech Republic, September 27, 2013. In Georg Krempl, Indre Zliobaite, Yin Wang, and Georg Forman (Eds.), (Online)Magdeburg University, 2013. URL
Correcting the Usage of the Hoeffding Inequality in Stream Mining. In Allan Tucker, Frank Höppner, Arno Siebes, and Stephen Swift (Eds.), Advances in Intelligent Data Analysis XII, (8207):298-309, Springer Berlin Heidelberg, 2013. URL
Advanced Topics on Data Stream Mining: Part II. Mining Multiple Streams. Bristol, UK, 24-28 Sept. 2012.
A hierarchical tree layout algorithm with an application to corporate management in a change process. Expert Systems with Applications, (39)15:12123-12130, 2012.
Tagungsband der 1. Doktorandentagung Magdeburger-Informatik-Tage (MIT 2012). In Georg Krempl, Claudia Krull, Frank Ortmeier, Eike Schallehn, and Sebastian Zug (Eds.), 2012. URL
Classification in Presence of Drift and Latency. In Myra Spiliopoulou, Haixun Wang, Diane Cook, Jian Pei, Wei Wang, Osmar Zaïane, and Xindong Wu (Eds.), Proceedings of the 11th IEEE International Conference on Data Mining Workshops (ICDMW 2011), IEEE, 2011.
The Algorithm APT to Classify in Concurrence of Latency and Drift. In João Gama, Elizabeth Bradley, and Jaakko Hollmén (Eds.), Advances in Intelligent Data Analysis X, (7014):222-233, Springer, 2011.
Drift Models and Classification in Presence of Latency and Drift. Proceedings of the Symposium Learning, Knowledge, Adaptation (LWA 2011) of the GI Special Interest Groups KDML, IR and WM., 65--72, September 2011.
Online Clustering of High-Dimensional Trajectories under Concept Drift. In Dimitrios Gunopulos, Thomas Hofmann, Donato Malerba, and Michalis Vazirgiannis (Eds.), Machine Learning and Knowledge Discovery in Databases, (6912):261-276, Springer Berlin Heidelberg, 2011. URL
Data-intensive, evidence-based decision making that builds on statistical machine learning approaches is becoming widely used through science, technology and commerce. While the number of applications using these approaches, as well as the volume, velocity and variability of data increases, the capacities of processing systems and of (human) supervisors remain limited. Furthermore, many applications operate in non-stationary environments that require fast reactions to new situations, even if little data or supervison is yet available. These challenges concern the whole data science workflow. Thus, my research is focused on the development and integration of efficient adaptive approaches that optimize the whole data science workflow, interact intelligently with their volatile environment, and consider constraints such as limited supervision, processing, or storage capacities.
This includes techniques for active, semi-supervised, and transfer learning, as well as adaptive online processing techniques, and methods for change detection, diagnosis and mining, for applications mainly in business, economics/social sciences, and neurosciences.
- Data Science, Applied Statistics, Machine Learning, Data Mining
- Development of novel approaches for evolving and streaming data,
- that handle constraints in supervision, processing, or storage capacities,
- by using temporal transfer, semi-supervised, and active machine learning techniques.
Selected Recent Research Topics
- Orange Labs France - Vincent Lemaire
- Leibniz Institute for Neurobiology Magdeburg - Matthias Deliano
- Department of Statistics and Operations Research at University of Graz
- Workshop and Tutorial on Active Learning: Applications, Foundations and Emerging Trends at iKNOW 2016
- Tutorial on Active Learning (and best paper) at iKNOW 2015
- Tutorial on Mining Multiple Threads of Streaming Data at PAKDD 2013
- Workshop Real-World Challenges for Data Stream Mining at ECML PKDD 2013
- Tutorial Advanced Topics on Data Stream Mining at ECML PKDD 2012
Teaching and Supervision
My teaching comprises several courses in the areas of data science, data mining and machine learning, as well as courses on foundations in topics such as information technology in organizations, statistics, operations research, combinatorial optimization, graph theory, and programming for non-computer-scientists.
I am currently supervising students on research projects (bachelor software projects and master team projects) and bachelor and master seminars on topics related to my research. Topics I'm most interested are on novel data science/machine learning approaches that use temporal transfer, semi-supervised, and active machine learning techniques in steamining and evolving data and applications in neurosciences, social sciences and economics/business. If you are interested, contact me via email.
Recent and Current Teaching
- Summerschool Learning Systems / Biocomputing (within the bachelor study profile of the same name)
- Research-Oriented Bacheor Module Human-Learner Interaction
- Master Module Data Mining in Changing Environments
- Information Technology in Organizations
- Bachelor and Master Seminars and Research Projects
For further details on currently offered courses, see LSF.