Modern industrial automation systems incorporate a variety of interconnected sensors and actuators that contribute to the generation of vast amounts of data. Although valuable insights for plant operators and engineers can be gained from such data sets, they often remain undiscovered due to the problem of applying machine learning algorithms in high-dimensional feature spaces. Feature selection is concerned with obtaining subsets of the original data, e.g. by eliminating highly correlated features, in order to speed up processing time and increase model performance with less inclination to overfitting. In terms of highdimensional data produced by automation systems, lots of dependencies between sensor measurements are already known to domain experts. By providing access to semantic data models for industrial data acquisition systems, we enable the explicit incorporation of such domain knowledge. In contrast to conventional techniques, this semantic feature selection approach can be carried out without looking at the actual data and facilitates an intuitive understanding of the learned models. In this paper we introduce two semantic-guided feature selection approaches for different data scenarios in industrial automation systems. We evaluate both approaches in a manufacturing use case and show competitive or even superior performance compared to conventional techniques.

No abstract

Smart factories are equipped with machines that can sense their manufacturing environments, interact with each other, and control production processes. Smooth operation of such factories requires that the machines and engineering personnel that conduct their monitoring and diagnostics share a detailed common industrial knowledge about the factory, e.g., in the form of knowledge graphs. Creation and maintenance of such knowledge is expensive and requires automation. In this work we show how machine learning that is specifically tailored towards industrial applications can help in knowledge graph completion. In particular, we show how knowledge completion can benefit from event logs that are common in smart factories. We evaluate this on the knowledge graph from a real worldinspired smart factory with encouraging results.

No abstract

Statistical learning of relations between entities is a popular approach to address the problem of missing data in Knowledge Graphs. In this work we study how relational learning can be enhanced with background of a special kind: event logs, that are sequences of entities that may occur in the graph. Events naturally appear in many important applications as background. We propose various embedding models that combine entities of a Knowledge Graph and event logs. Our evaluation shows that our approach outperforms state-of-the-art baselines on real-world manufacturing and road traffic Knowledge Graphs, as well as in a controlled scenario that mimics manufacturing processes. HE Knowledge Graph zero-shot entity ( ) LE LE Board Jam HE ( ) LE LE Coil Jam LE ( )