CC BY-NC-ND 4.0Klau, DennisDennisKlauZöller, Marc-AndréMarc-AndréZöllerTutschku, Christian KlausChristian KlausTutschku2024-01-152024-01-152023https://publica.fraunhofer.de/handle/publica/458782https://doi.org/10.24406/publica-243710.48550/arXiv.2310.0423810.24406/publica-2437Quantum Computing (QC) is becoming an increasingly promising technology for modern computation, especially in the field of data driven approaches like simulation and machine learning. With the high momentum of research and development of new hard- and software, QC holds a big promise in redefining many state-of-the-art computation approaches today. On the other hand, the nowadays well-established field of machine learning (ML) faces challenges like the discrepancy of demand by industry and availability of ML experts, reproducibility, and efficiency in prototyping. To overcome some of these issues, several frameworks have been created for automating the process of pipeline construction, data preprocessing, model training and hyperparameter optimization (HPO), many of them open source. In most cases, these Automated Machine Learning (AutoML) frameworks implement a fixed subset of known approaches and algorithms, or encapsulate an established ML backend, that defines the available algorithms. This work describes the selection approach and analysis of existing AutoML frameworks regarding their capability of a) incorporating Quantum Machine Learning (QML) algorithms into this automated solving approach of the AutoML framing and b) solving a set of industrial use-cases with different ML problem types by benchmarking their most important characteristics. For that, available open-source tools are condensed into a market overview and suitable frameworks are systematically selected on a multi-phase, multi-criteria approach. This is done by considering software selection approaches [1], as well as in terms of the technical perspective of AutoML [2, 3]. The requirements for the framework selection are divided into hard and soft criteria regarding their software and ML attributes. Additionally, a classification of AutoML frameworks is made into high- and low-level types, inspired by the findings of [4]. Finally, we select Ray and AutoGluon as the suitable low- and high-level frameworks respectively, as they fulfil all requirements sufficiently and received the best evaluation feedback during the use-case study. Based on those findings, we build an extended Automated Quantum Machine Learning (AutoQML) framework with QC-specific pipeline steps and decision characteristics for hardware and software constraints.enpaper