No abstract

No abstract

Finding redundant requirements or semantically similar ones in previous projects is a very time-consuming task in engineering design, especially with multilingual data. Due to modern NLP it is possible to automate such tasks. In this paper we compared different multilingual embeddings models to see which of them is the most suitable to find similar requirements in English and German. The comparison was done for both in-domain data (requirements pairs) and out-of-domain data (general sentence pairs). The most suitable model were sentence embeddings learnt with knowledge distillation.

Companies are progressively gathering data within the digitalization of production processes. By actively using these production data sets operational processes can be improved, hence empowering businesses to compete with other corporations. One way to achieve this is to use data from production processes to develop and offer smart services that enable companies to continuously improve and to become more efficient. In this paper, the authors present an industrial use case of how machine learning can be implemented into smart services in production processes to decrease the time for resolving upcoming issues in manufacturing. The implementation is carried out by using an assistance system that aids a team which attends to problems in the assembling of turbines. Therefore, the authors have analyzed the assembly problems from an issue management system that the team had to resolve. Subsequently three different approaches based upon natural language processing, regression and clustering were selected. This paper also presents the development and evaluation of the assistance system.

No abstract