Marc Zimmermann

Hugo Yèche, Rita Kuznetsova, Marc Zimmermann et al.

The recent success of machine learning methods applied to time series collected from Intensive Care Units (ICU) exposes the lack of standardized machine learning benchmarks for developing and comparing such methods. While raw datasets, such as MIMIC-IV or eICU, can be freely accessed on Physionet, the choice of tasks and pre-processing is often chosen ad-hoc for each publication, limiting comparability across publications. In this work, we aim to improve this situation by providing a benchmark covering a large spectrum of ICU-related tasks. Using the HiRID dataset, we define multiple clinically relevant tasks in collaboration with clinicians. In addition, we provide a reproducible end-to-end pipeline to construct both data and labels. Finally, we provide an in-depth analysis of current state-of-the-art sequence modeling methods, highlighting some limitations of deep learning approaches for this type of data. With this benchmark, we hope to give the research community the possibility of a fair comparison of their work.

Daniel Fraunholz, Marc Zimmermann, Hans D. Schotten

Since honeypots first appeared as an advanced network security concept they suffer from poor deployment and maintenance strategies. State-of-the-Art deployment is a manual process in which the honeypot needs to be configured and maintained by a network administrator. In this paper we present a method for a dynamic honeypot configuration, deployment and maintenance strategy based on machine learning techniques. Our method features an identification mechanism for machines and devices in a network. These entities are analysed and clustered. Based on the clusters, honeypots are intelligently deployed in the network. The proposed method needs no configuration and maintenance and is therefore a major advantage for the honeypot technology in modern network security.

Simon Duque Anton, Daniel Fraunholz, Christoph Lipps et al.

Since the early 1960, industrial process control has been applied by electric systems. In the mid 1970's, the term SCADA emerged, describing the automated control and data acquisition. Since most industrial and automation networks were physically isolated, security was not an issue. This changed, when in the early 2000's industrial networks were opened to the public internet. The reasons were manifold. Increased interconnectivity led to more productivity, simplicity and ease of use. It decreased the configuration overhead and downtimes for system adjustments. However, it also led to an abundance of new attack vectors. In recent time, there has been a remarkable amount of attacks on industrial companies and infrastructures. In this paper, known attacks on industrial systems are analysed. This is done by investigating the exploits that are available on public sources. The different types of attacks and their points of entry are reviewed in this paper. Trends in exploitation as well as targeted attack campaigns against industrial enterprises are introduced.

Stephanie L. Hyland, Martin Faltys, Matthias Hüser et al.

Intensive care clinicians are presented with large quantities of patient information and measurements from a multitude of monitoring systems. The limited ability of humans to process such complex information hinders physicians to readily recognize and act on early signs of patient deterioration. We used machine learning to develop an early warning system for circulatory failure based on a high-resolution ICU database with 240 patient years of data. This automatic system predicts 90.0% of circulatory failure events (prevalence 3.1%), with 81.8% identified more than two hours in advance, resulting in an area under the receiver operating characteristic curve of 94.0% and area under the precision-recall curve of 63.0%. The model was externally validated in a large independent patient cohort.

Amina Fellan, Christian Schellenberger, Marc Zimmermann et al.

Within the context of Industry 4.0, mobile robot systems such as automated guided vehicles (AGVs) and unmanned aerial vehicles (UAVs) are one of the major areas challenging current communication and localization technologies. Due to stringent requirements on latency and reliability, several of the existing solutions are not capable of meeting the performance required by industrial automation applications. Additionally, the disparity in types and applications of unmanned vehicle (UV) calls for more flexible communication technologies in order to address their specific requirements. In this paper, we propose several use cases for UVs within the context of Industry 4.0 and consider their respective requirements. We also identify wireless technologies that support the deployment of UVs as envisioned in Industry 4.0 scenarios.