This overview describes our project still in development phase.
The Spot Inconsistencies in Tax-related Transactional Data project was successfully launched on 1 January 2018. The aim of the “SPOTTeD” project is the automated analysis of transaction data for the recognition and visualization of transactions that have been wrongly classified for tax purposes, tax-critical transactions and the optimization of linked business processes with regard to potential tax savings. To this end, the potentials of deep learning methods for detecting anomalies within transaction data and of reinforcement learning methods for optimization according to control criteria are investigated. Responsible for this project is our researcher Johannes Lahann and the expected duration is 18 months.
The Safety Expert 4.0 project was successfully launched on 1 June 2017. SiFa 4.0 aims to automate the manual processes in Hazardous Substance Management as far as possible and thus reduce costs by 80%. Particular emphasis is placed on the search and evaluation of substitutes (substitution testing), which is a complex task for two million safety data sheets in Europe. The automated substitution check ensures on the one hand that workers are not exposed to unnecessary risks and on the other hand that previously hidden cost savings potentials in production are visible.Contact person at DFKI for this project is Patrick Lübbecke and the expected duration of this project is 24 months.
The project Process-Based Blockchain for the monitoring of cross-company business processes was successfully launched on January 1, 2018. In the ProcessChain project, the possibilities and consequences of this technology for monitoring cross-company business processes are to be tested in practice. The digitization and networking of all business processes has been a stable trend for years, affecting not only internal processes within the company, but more recently especially cross-company processes. The trust of the actors involved creates the basis for innovation and efficiency of cooperation. Blockchain technology as the technical basis of the cryptocurrency bitcoin opens up a wide range of innovations. In the proposed project, this technology will be used to implement a process-based block chain for monitoring cross-company business processes. The ProcessChain platform implemented in the project enables cross-company processing of business processes without the use of a central control authority, which all partners involved have to trust. The platform’s considerable innovation potential is piloted and evaluated on selected cross-company demonstration processes from various industries such as service management and supply chain management.The responsible contact persons for this project at DFKI are Philip Hake and Filip Fatz.
The Data Sovereignty Manager project was successfully launched on 1 January 2018. The Data Sovereignty Manager (DaSoMan) project will give users of web services complete control over their personal data. At the same time, providers and developers of web services are to be provided with simple mechanisms to integrate helpful analytic functionalities and still guarantee data security and transparency for their users. With the DaSoMan, providers can efficiently obtain legal certainty regarding the soon to be valid EU basic data protection regulation and, if necessary, have their application certified efficiently in the aspect of data protection and data security (advantage of trust for users).The responsible contact persons for this project at DFKI are Andreas Emrich and Michael Frey.
Due to the continuously increasing number of production orders with high number of units as well as the simultaneous demand for the production of orders with small numbers of units, the manufacturing industry is particularly affected by dynamic changes in the industry 4.0. Through the introduction of concepts such as the continuous networking of production machines to cyber-physical systems, the basic infrastructure necessary to implement this dynamic was created. Various sensors within production machines make enormous amounts of data accessible to a variety of analysis purposes, depending on the size of the production lines.
In order to implement the necessary flexibility also within the coarse and detailed planning of the production, complex dependencies between different information on order situation, personnel and machine availability must be used. So far, production planning in real-time is not possible or can not be adapted, but is carried out with a time offset from hours to days before the actual production.
In the project ProPlanE, an analysis platform is designed and implemented to integrate data from different systems and to support a production planning in real time with the aid of process mining. Continuous monitoring of expiring processes with respect to the dimensions of time, resources and costs allows a continuous adjustment of the production planning as well as a flexible handling of short-term changes in the production process.
Tha project was launched in January 2017 and is expected to be completed at the end of July 2018.