Companies are currently very keen to harness their data capital for future-proof smart products and sustainable value creation. With the help of artificial intelligence (AI), they are analyzing their large volumes of data (Big Data) and trying to create added value from their data. However, medium-sized companies, especially SMEs, have much smaller amounts of data available compared to large platform companies, so-called Small Data. These often sparse data sets (Scarce Data), such as lab data, machine performance data, people knowledge (reports), and equipment usage data, have high value for the companies, as they contain information about their products and processes, as well as their performance and innovation potential. The project is developing an AI solution consisting of information fusion and interpretable AI in modular form, which also supports further data collection, e.g., in the lab, through a feedback loop. The solutions will be transferred to application in the consortium, pursuing various economic goals of the partners, such as product and process design for plant breeding, product and process design for separators, and diagnostic support for networked systems, among others.

Many innovations of the companies from OWL originate from sustainable ideas, which are created or developed by or with experiments, laboratory tests and by prototypes in real use. In the process, targeted data are collected that serve to design, evaluate and test the innovations. A collection of these data is only possible to a limited extent due to high financial and time expenditures or a lack of reproducibility, which is why in most cases the data is Scarce Data (jewels of the companies), in which, however, important information about the behavior and performance of these innovations is contained. In contrast to Big Data, Scarce Data stands for few and at the same time precise data, which can occur in different formats (heterogeneous data). The AI4ScaDa project exploits this field of innovation and explores specific methods of ML to profitably take advantage of the small and often heterogeneous data. Challenges are that AI systems depend heavily on the quality and quantity of data. Biases present in the training data are reflected in the inference after the learning phase. Although AI systems can learn static correlations from enormous data, they have no understanding of causal relationships. Where very sparse data is available, a large number of AI methods have difficulty keeping up with human expertise.

The AI4ScaDa project pursues an economic and benefit-oriented as well as a methodological objective. The focus is on use cases that are characterized by scarce data and heterogeneous data sources. Use cases of the participating partner companies SAATEN-UNION BIOTEC GmbH, GEA Westfalia Separator Group GmbH and Miele & Cie. KG will confirm the benefit and transferability of the developed solution. The focus is on (i) a product and process design for plant breeding, (ii) a product and process design for separators and (iii) a diagnostic support for networked systems. The goal of all applications is to profitably use laboratory data, coupled with other data sources, for innovative products and services that are fit for the future. To this end, the project pursues an overarching methodological goal by developing an AI solution consisting of an information fusion and interpretable AI in a modular and generalized form, which also supports data collection, e.g. in the laboratory, by means of a feedback loop.

The added value of AI4ScaDa lies in the novel use of the data capital of Scarce Data for all types of companies, especially for SMEs. Especially in data, which is only available to a limited extent due to cost-intensive procedures, measurements, experiments, etc., there is an enormous potential for innovation. The AI4ScaDa project uses this potential and proves, based on the applications in the participating companies, that scarce data in combination with fusion-based AI delivers sustainable added value for all types of companies. AI4ScaDa thus represents an essential component for achieving technological leadership in the field of Small Data technologies. For SU BIOTEC, AI4ScaDa generates an economic benefit by optimizing and simplifying the processes for the production of new starting materials for plant breeding based on Scarce Data (data from e.g. laboratory and greenhouse). This innovation contributes significantly to environmental sustainability for humans and the environment by enabling plants to adapt more quickly to climatic changes, thereby positively impacting the ecosystem. For GEA, laboratory data is efficiently harnessed, leading to resource-efficient adaptation of machines (separators) to customer requirements. AI4ScaDa as a service helps to conserve material and reduce energy consumption of machines. Miele will use the fusion-based AI workflow to enrich lab data from, for example, product trials with equipment usage data to achieve diagnostic support for hard-to-identify situations, which simultaneously reduces customer service calls and extends product life cycles. By using AI4ScaDa, companies are enabled to achieve a higher degree of autonomy in the use of scarce data and thus address environmental and economic sustainability aspects.