Smart Data Innovation Challenges

The aim of the SDIC micro-projects is to achieve a short period of time between the idea and the implementation of AI research. From August 2021, the BMBF will fund up to 7 additional AI micro-projects at DFKI, Forschungszentrum Jülich, Fraunhofer Gesellschaft and KIT, each with a maximum of 6 person-months of industrial data in the Smart Data Innovation Lab (SDIL). Funding is given to the best project ideas with a clear reference to industrial data with the involvement of companies (or the public sector) as associated application partners and data providers. Several projects are to be funded in each of the following priority areas:
AI micro-projects on spatial data or graphs
Micro-projects are to be prioritized, which analyze relationships on multidimensional, spatial data and networks (such as hyperspectral remote observation data or heterogeneously linked data from knowledge graphs) using AI. In particular, new types of learning architectures are to be promoted, which model the relationship between data points and beyond simple sequential patterns or 2D conversions. The focus is on data relevant to industrial practice that cannot be easily applied using classic ML methods for tables, bitmaps or sequences.
AI micro-projects on computationally intensive, data-saving learning
The availability of large, suitable data sets is often a major hurdle for the application of models such as deep neural networks. Priority should be given to projects that use the computing power of the SDIL to adapt models or data to questions that would not be addressable with classic data-driven machine learning approaches. However, it is also important here to have a representative amount of industrial data to validate the approaches.
Open-topic AI micro-projects throughout the model life cycle
In addition, new research impulses with high application potential are to be promoted. The focus should be on the practical implementation of AI on operational data. The aim is to show how models learned offline can be optimized for execution within the SDIL and in the connected cloud solutions. For this purpose, the SDIL provides in particular cloud services from the platform partners (IBM, SAP and Software AG): . Aspects of the model life cycle from learning to monitoring and continuous validation of models should be viewed in terms of utilization.
Update: Extension of the tender by one week until July 12, 2021, due to a technical problem with the submission email address (If you do not receive a confirmation of the submission, please call Till Riedel on +49 721 608 41706)


In addition to the scientific quality, the innovative use of industrial data sources and a clear scientific and economic exploitation plan, the profitable use of computing resources in the SDIL is relevant for the selection.


In particular, the integration of data and issues from small and medium-sized companies that benefit from the cooperation as unsupported partners is desired.
Applicants will receive feedback on the success of the review within 4 weeks of the submission deadline and can start their research as soon as possible.
The fully completed application for funding under SDI-C Call 4 must be sent to the office of the consortium leader by July 5, 2021 (preferably in docx format to Personnel costs of the researchers involved in the project at DFKI, Forschungszentrum Jülich, Fraunhofer Gesellschaft and KIT are eligible. Interested companies can apply at an early stage to even without an academic partner if they are looking for a solution to interesting problems with your data. In addition, there will be an online information event with networking opportunities on June 10th, 2021 at 11:30 a.m.
calendar (1)
The SDIL offers a legally and technically secure framework for the processing of the data within the scope of the requested micro-project (see sample data usage contract for micro-projects in Appendix 2). The individually adaptable sample cooperation agreement (Annex 3) for micro-projects, which the research and industrial partners and KIT as the operator of the infrastructure agree on before the start of the project in accordance with the overall cooperation agreement, is attached. Within this contract, the partners clearly regulate the respective rights of use to the data and the results, so that the interests of all parties are protected. Examples of successful cooperation projects with industrial partners can be found at:
Computer and cloud resources in the SDIL are made available to the micro-projects free of charge. The SDIL is also happy to advise applicants in advance. Technical and organizational requirements of the respective micro-project on the infrastructure can be clarified in advance with the operations team:
Once the funds have been released by the project management company and a cooperation agreement with the data provider has been concluded, the project can start immediately. Funded micro-projects in this round should be completed by March 2022 at the latest. The results will be presented to the public in events and on the website ( ). In particular, the projects should lead to a timely joint exploitation of the results with the participating company.
The release of the funds for the best project ideas is requested after a positive assessment on the basis of a joint steering committee decision by the lead agency of the respective partner at the BMBF and is finally carried out after assessment by the funding agency as part of the existing project, so that the funds are available within a few weeks can stand! In addition, you can use the templates regardless of the deadline for submitting applications to apply for “unsupported” micro-projects, for which the use of the infrastructure is then applied for.