Our research on intelligent information systems (I2S) is concerned with the application of Artificial Intelligence (AI), services computing and semantic technologies for innovative means of intelligent data analysis and service coordination in smart information systems in the Internet.
Intelligent information systems in the future Internet of Things and Services are envisioned to provide users with intelligently assisted real-time access to relevant data and service coordination anywhere, anytime, on any device, to anyone, safe and secure. Agent technologies for individual or joint AI action planning and learning, inter-agent communication, negotiation and coordination can be used to implement autonomous, rational, proactive, adaptive and collaborative behaviour of such systems with intelligent agents. Services computing enables intelligent agents to provide and coordinate (discover, select, negotiate, compose, execute) their own and other services more flexibly and efficiently in different IT environments. Semantic technologies enable the semantic interoperation of distributed heterogeneous data, services and process models by these agents for intelligent information systems in support of industry and business, as well as our private and social life.
The I2S research team of the DFKI-ASR department has long-standing, internationally recognized expertise and competencies in the following research areas:
- Semantic search and composition planning of services in dynamic, heterogeneous and distributed environments
- Automated negotiation of charged data and services (bargaining, coalition formation, auctions)
- Semantic data stream analysis with uncertainty and semantic explanations to human users
- Hybrid deep learning for scene interpretation and action planning by autonomous agents
The I2S team offers innovative and implemented technologies for
- Service coordination
- Semantic data analysis
- Negotiation mechanisms
- Process optimisation
Selection of our open-source and partly awarded software tools and benchmarks.
Our technologies for intelligent data analysis and service coordination have been applied to various industrial use cases:
- Production planning and logistics in agriculture and forestry
- Process optimisation in automotive and metal processing
- Intelligent condition monitoring / predictive maintenance
- Semantic business process management in smart retail
- Smart event management and social media analysis
- Smart decentralized power regulation in micro-grids
- Mobile medical assistance services in health care
- Adaptive navigation in autonomous driving
Selected intelligent information systems of our team have also been demonstrated at the CEBIT, the Hannover Industry Fair (HIM) and the International Consumer Electronics Show (CES).
Our research and development has been conducted in many projects funded by the European Commission, national government and industry. The team contributions to selected research projects are as follows:
- REACT (BMB+F; 2017 – 2020): Autonomous Driving – Learning and Simulation Environment for Critical Traffic Scenarios. Team contributions: Hybrid deep learning for scene interpretation and navigation planning by autonomous cars and pedestrians in critical traffic scenarios. Simulation of adaptive (hybrid deep learning based) behavior of autonomous agents (cars, pedestrians) in virtual 3D environment with openDS.
- CREMA (EC H2020 Factory of the Future, RIA; 2015 – 2017): Cloud-Based Rapid Elastic Manufacturing. Team contributions: Optimization of semantic service-based business process models in BPMN at design time and runtime; Largest publicly available manufacturing ontology CDM-Core with focus on machinery maintenance and automotive exhaust production.
- INVERSIV (BMB+F; 2014 – 2017): Integrated Verification, Simulation and Visualization for Industrial Applications. Team contributions: Semantic and probabilistic sensor data stream analysis for fault detection and diagnosis; Integration with model-based prognosis and worker support; Parallel and distributed solutions of NP-hard manufacturing problems.
- ICM-Hydraulic (HYDAC GmbH; 2013 – 2015): Intelligent Condition Monitoring of Hydraulic Machines. Team contributions: Combined statistical and hybrid semantic sensor data stream analysis for highly accurate fault detection and diagnosis of hydraulic machines; Automated generation of semantic explanation of diagnosis result to the human operator; Application to condition monitoring of wind turbines.
- SocialSensor (EU FP7, IP; 2011 – 2014): Real-Time Multimedia Content Search and Analysis in the Social Web. Team contributions: Semantic middleware; Automated semantic search and composition planning of media servicess; Mobile (Android) application MyMedia for semantic end-to-end search and streaming of videos and live recordings.
Selected publications on different research results of the team and their application in different industries are listed below:
- Pusse, F.; Klusch, M. (2019): Hybrid Online POMDP Planning and Deep Reinforcement Learning for Safer Self-Driving Cars. Proc. 30th IEEE International Intelligent Vehicles Symposium (IV), IEEE.
- Mazzola, L.; Kapahnke, P.; Klusch, M. (2017): Pattern-Based Semantic Composition of Optimal Process Service Plans with ODERU. Proc. 19th ACM International Conference on Information Integration and Web-based Applications and Services (iiWAS); Salzburg, Austria, ACM.
- Klusch, M.; Meshram, A.; Schuetze, A.; Helwig, N. (2015): ICM-Hydraulic: Semantics-Empowered Condition Monitoring of Hydraulic Machines. Proc. 11th International Conference on Semantic Systems (SEMANTiCS); Vienna, Austria; ACM
- Mihailescu, R.; Ossowski, S.; Klusch, M. (2016): eCOOP: Applying Dynamic Coalition Formation to the Power Regulation Problem in Smart Grids. Computational Intelligence, Wiley.
- Kahl, G.; Klusch, M.; Zinnikus, I.; Schimmelpfennig, J.; Zapp, M. (2015): ADIGE: Semantic Business Process Management for Smart Retail Environments. Proc. 17th ACM International Conference on Information Integration and Web-based Applications and Services (iiWAS), Brussels, Belgium; ACM.
- Klusch, M.; Kapahnke, P.; Cao, X.; Rainer, B.; Timmerer, C.; Mangold, S. (2014): MyMedia: Mobile Semantic Peer-to-Peer Video Search and Live Streaming. Proc. 11th ACM International Conference on Mobile and Ubiquitous Systems (Mobiquitous); London, UK; ACM Press
- Mazzola, L.; Kapahnke, P.; Vujic, M.; Klusch, M. (2016): CDM-Core: A Manufacturing Domain Ontology in OWL2 for Production and Maintenance. Proc. 8th International Conference on Knowledge Engineering and Ontology Development (KEOD), Porto, Portugal.
- Klusch, M.; Kapahnke, P.; Schulte, S.; Lecue, F.; Bernstein, A. (2016): Semantic Web Service Search: A Brief Survey. Kuenstliche Intelligenz, 30(2), Springer.
- Cao, X.; Kapahnke, P.; Klusch, M. (2015): SPSC: Efficient Composition of Semantic Services in Unstructured P2P Networks. Proc. 12th Extended Semantic Web Conference (ESWC); LNCS 9088, Springer
- Degenbaeva, C.; Klusch, M. (2015): Critical Node Detection Problem Solving on GPU and in the Cloud. Proc. 17th IEEE International Conference on High-Performance Computing and Communications (HPCC); New York, USA
- Klusch, M. (2014): Service Discovery. In: Alhajj, R.; Rokne, J. (eds.), Encyclopedia of Social Networks and Mining (ESNAM), Springer.
PD Dr. Matthias Klusch (Team leader)
I2S Research Team: http://www.dfki.de/~klusch/i2s