The main challenges are discussed together with the lessons learned from past and ongoing research along the development cycle of machine learning systems. This will be done by taking into account intrinsic conditions of nowadays deep learning models, data and software quality issues and human-centered artificial intelligence (AI) postulates, including confidentiality and ethical aspects. The analysis outlines a fundamental theory-practice gap which superimposes the challenges of AI system engineering at the level of data quality assurance, model building, software engineering and deployment. The aim of this paper is to pinpoint research topics to explore approaches to address these challenges.

Architecture decision records (ADRs) have been proposed as a resource-efficient means for capturing architectural design decisions (ADDs), and have received attention not only from researchers but also from practitioners. We conducted a mining software repositories (MSR) study, in which we analyzed the use of ADRs in open source repositories at GitHub. Our results show that the adoption of ADRs is still low, although the number of repositories using ADRs is increasing every year. About 50% of all repositories with ADRs contain just one to five ADRs suggesting that the concept has been tried but not yet definitively adopted. In repositories that use ADRs more systematically, we observed that recording decisions is a team activity conducted by two or more users over a longer period of time. In most repositories the template proposed by Michael Nygrad is used. We, finally, provide an interpretation of the obtained results and discuss open future research challenges by elaborating on implications of the study’s findings as well as on recommendations on how to further increase the adoption of ADRs.

Supporting Customizable Business Process Models Using Graph Transformation Rules: 10.4018/IJISMD.2016070103: Business Process customization is an active research area in the process management field, dealing with variations/commonalities among processes of a given