Best Student Paper Award at 23th ICEIS for Cluster-Team
The falling incidence means a return to the usual physical conference operations is already close, but the 23rd ICEIS still had to take place from the home office this year and not, as originally planned, in Prague. For the Cluster of Excellence, the virtual International Conference on Enterprise Information from provided a perfect framework to take up and represent the subject of the Internet of Production. The conference, this year taking place April 26-28, brings together researchers in the field of application of information systems and builds a bridge between computer science theory and practical application, just like the research of the team consisting of Johannes Lipp, Siyabend Sakik, Moritz Kröger (LLT) and Stefan Decker (FIT) from the Cluster. This team of researchers present at the conference received the prestigious Best Student Paper Award for their presented work in the closing session. We asked Johannes Lipp from Fraunhofer FIT about the conference and the team contribution.
How did this year's ICEIS go from home office?
Because of Covid-19, the conference did not take place in Prague as usual, but at home in the home office, organizationally it was regulated in such a way that papers, presentations, posters and other materials were recorded in advance and loaded onto prepared websites for backup purposes to be able to use in case someone had technical problems during the sessions. We were Paper # 190 at the conference and wrote it together with the candidate following a bachelor thesis, which was a cooperation between Fraunhofer FIT and the LLT of RWTH Aachen University. The title of our paper is "LISSU: Integrating Semantic Web Concepts Into SOA Frameworks" On Wednesday evening we received the Best Student Paper Award in the closing session, a pretty big honor!
How can your team's contribution be briefly summarized?
The context of our work is the ultrashort pulse laser process, which I always like to refer to as "backward 3D printing". Very short, high-energy laser pulses gradually remove material. Background: The LLT does not buy finished machines, but puts them together in a modular fashion from components. The syntax (form) is automatically checked to ensure that the components that are frequently exchanged work together perfectly. This ensures, for example, that the motion module receives two numbers for the X and Y values. Our focus is on the interaction of this syntax and semantics, i.e. the meaning of the data. The problem manifests itself in the fact that with this automatic check only the form of the data is transmitted, but not its meaning (semantics). The following scenario often occurs: When a call is made (X = 5, Y = 5), the old movement module moves the laser to 5 millimeters, while a new one moves to 5 centimeters. One can imagine possible critical consequences ... Our approach helps prevent such problems. To extend the syntactic validation by a semantic one, which describes the units of all interfaces formally with the help of ontologies and checks them before each communication. How do we do that? It's about the implementation of small programs in Python, which expand all components of the system. Our approach is downward compatible, so it works with both extended and "normal" components. We have also implemented an automatic conversion of units. We have already tested our approach in real application scenarios and demonstrated the functionality of our approach at the LLT by preventing (faulty) communication in the case of incompatible units or by automatically converting this if possible.
What went particularly well about the conference and the collaboration that ultimately led to the committee award?
The collaboration with the student Siyabend Sakik went really well, on whose bachelor thesis we based this paper. Also communication with the organizers in advance who The submission and review process and the technical solutions of the conference were very commendable. On the one hand, it was worth mentioning the possibility of putting together your own personalized conference plan from the program overview and importing it into your own calendar, e.g. Outlook. The detailed program information was also helpful: for each paper, workshop and poster there was a detail page on which you can view all content including the recorded presentation, join the zoom session or discuss the content in a forum. There was also a Slack chat at the same time, which enables personal exchange between the participants.
In your opinion, what else can be optimized in terms of conference operations?
Since the conference covers many topics (e.g. data integration, machine learning, semantics and more), the thematic proximity of the papers in a session sometimes leaves a lot to be desired. The numerous communication options offered were a good idea, but were not used enough and the poster session, which before Corona mostly corresponded to a "lively bazaar", was unfortunately only a series of 1-minute videos in the online version without any real exchange. In addition, the quality of the internet, video and sound was very poor for some participants. Intercontinental meetings in particular still seem problematic.
Now that the prospect of a normalization of the Corona situation is slowly emerging, what are you most looking forward to when there will be physical conferences again?
The "person-to-person" exchange, which is unfortunately not satisfactory in the virtual world. But also to be able to use more soft skills again in lectures and to see them used, pre-recorded videos are no comparison to really being on stage in front of a large audience. At ICEIS, only around 60% of the lectures were given live, the rest by video recording. Of course, I'm also looking forward to cultural experiences, sights and travel itself again.