Partner

AviComp Controls GmbH

www.avicomp.com

The aim of SAKE is to design, implement and evaluate a platform which will successfully integrate and process large and diverse event data streams with particular emphasis on the requirements of engineering and IT monitoring systems. AviComp Controls are an applications partner of this subproject who supply the specially developed CCM-221 vibration monitoring system for compressor plants. Their most important target technical innovations are (1) the development of an algorithm to provide users with early recognition of aerodynamic instabilities, (2) the early detection of impending maintenance issues, (3) remote access to the CCM-221 in order to carry out long-term data base based measurements and analysis of machine conditions and breakdown prognosis and (4) adjustment and integration to the SAKE platform. In addition to these specific goals AviComp are also actively involved with other working packets within the framework of the subproject. An important aspect of this is the support they provide during the design of user requirements, the design and prototype conversion of the predictive analysis of error descriptions and a user study using real system data.

Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS

www.iais.fraunhofer.de

The Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS is one of the most prominent research institutes specializing in areas such as Data Science, Machine Learning, Information Retrieval and semantic technologies. The Fraunhofer IAIS brings the following expertise into the SAKE project: (1) development of semantic Web ontologies, (2) data analysis and visualization as used for searching for local patterns and interactive visualization, (3) event data analysis for processing complex events in Big Data architectures.

Heidelberger Druckmaschinen AG

www.heidelberg.com

The Heidelberger Druckmaschinen AG have already developed an expertise with online monitoring and the analysis of printing machines through their Remote Diagnosis system. Their Remote Services department uses systems specifically designed to check the log files of machines for known errors. They have also developed a prototype system aimed at the early detection of errors. It works by using specific indicators within basic set parameters and emits an advance warning in sufficient time to avoid unplanned machine breakdowns. Considering, however, the size and complexity of future data volumes and the many possibilities offered by Big Data technologies their current methods will at some point no longer be sufficient and it is for this reason that the prototype will continue to be developed within the Web-Based Remote Monitoring project.

Automatically generated algorithms (EPN – Event Processing Network) describe the causal connections which use a trigger point (error time) within the log file data of a printing system.

Heidelberg have also introduced to the project application-specific cases, printing systems' mass data and expertise in domain-specific data analysis. Verification of the analytic results takes place in pilot markets within the global Heidelberg service organisation.

Ontos GmbH

www.ontos.com

The clear focus of Ontos within the SAKE project is the conceptual design and prototypical implementation of the semantic layer that all other SAKE platform components are based on. Prerequisites for this are the pre-filtration and semantification of the input data in combination with the scalable, distributed storage of the generated RDF data. Ontos also continue to be involved in intelligent information visualization and the integration of the SAKE platform with other application partners.

University Leipzig

aksw.org 

The AKSW research group is an international group of experts specializing in semantic technology and Big Data fields. The group won the prestigious Best Research Paper Award at the ESWC 2013 (focussing on "Big Data and Semantics") based on their work on the scalable and distributed processing of similarity recognition processes. They also won the ISWC 2013 Big Data Challenge for their work on processing Big Linked Data as well as the I-Semantics Challenge for their work on the management of the world's second largest RDF knowledge base.

The AKSW research group enriches SAKE with six technological approaches: (1) The LIMES framework enables the computation of similarities (of events) in a scalable and time-efficient way by using quantity operators to combine dedicated time-efficient algorithms. Importantly, LIMES also implements supervised and unsupervised machine learning procedures aimed at further understanding the measurement of similarities. LIMES is the fastest Link Discovery Framework currently in existence anywhere in the world and has received several prizes, most notably the Best Research Paper Awards at the ISWC 2011 and the ESWC 2013. (2) The DL-Learner tool makes it possible to learn class expressions as time efficiently as possible by using supervised learning processes based on refinement operators. The algorithms implemented in this tool have also been awarded several prizes, including two Best Student Research Paper Awards at the ILP 2007. (3,4,5) DAW, Hibiscus and TopFed are federated query engines developed by the AKSW which allow the doublet-free and hypergraph-based processing of federated SPARQL queries. These technologies will form the basis of the federated query layer as well as AP3. (6) SPARQL2NL is a verbalization framework for RDF and SPARQL. The framework has already been used in question answering engines such as TBSL and benchmark frameworks such as BioASQ and will form the basis of verbalizing learning results in SAKE. As members of the computer center's research department based at the University of Leipzig, the AKSW have a petaflop server at their disposal for Big Data experiments. Because of this access the AKSW is primarily responsible for converting scalable query processes for the supervised and unsupervised scalable understanding of event descriptions. ULEI supports the Ontos GmbH with benchmarking the expansion of OntoQuad as well as the USU AG and their project to convert root cause analysis algorithms.

USU Software AG

www.usu.de 

USU Software AG is the parent company of the USU Group and is listed in the Prime Standard section of the German Stock Exchange. The company was founded in 1977 and offers its customers strategies, consultation, software solutions and services based on an all-encompassing combination of knowledge and experience of corporate processes and applications. USU is Europe's largest supplier of IT and knowledge management solutions in their respective business segments. Their Valuemation series of products gives the company a high quality portfolio with which to support ITIL-based IT and service management. USU use KnowledgeCenter technology alongside their own active knowledge databases to answer over 25 million separate queries to their call and service center every year – all of them handled efficiently and to a high standard. The company's portfolio is complemented by the USU Lifebelt self-service solution and the USU KnowledgeScout interactive advisory system for technical field staff. Additional core skills include process and system integration based on service orientated architectures (SOA). The USU research facility develops industrial Big Data systems designed to continually monitor and analyze sensor data as well as related automated service systems. USU is responsible for the entire coordination and integration of the SAKE project with an additional focus on data processing, user interfaces, usability and use cases.