At the IEEE AIoT2024 conference, Tianheng Ling, Chao Qian and Gregor Schiele won the Best Paper Award for their work on the efficient implementation of integer-only quantised Transformers on embedded FPGAs.

In May, a new research project was launched, the consortium of which we are leading: TransfAIr.

In this project, which is funded by the Federal Ministry of Education and Research, we are working together with the Chair of Distributed Systems of Prof. Dr.-Ing. Torben Weis and the companies HARTING, Krohne Messtechnik and Kuntze Instruments to develop and test transfer approaches for artificial intelligence in industry.

The methodological starting points of the project are (1) approaches for transferring learned knowledge to new problems or models (so-called transfer learning) and (2) approaches for the automated optimization of AI models (so-called AutoML). These are to be further developed and made usable for industrial users. For the evaluation, we use existing AI solutions and training data from our industrial partners.

In the renowned Springer journal CCF Transactions on Pervasive Computing and Interaction, our article Elastic AI: system support for adaptive machine learning in pervasive computing systems has recently been ranked 12th in the list of most downloaded articles.

CCF Transactions on Pervasive Computing and Interaction is a platform that publishes high-quality scientific articles on all aspects of human-computer interaction and pervasive/ubiquitous computing. Our article was written by Christopher Cichiwskyj, Stephan Schmeißer, Chao Qian, Lukas Einhaus, Christopher Ringhofer and Gregor Schiele.

We are very pleased that our research article has now been downloaded over 3264 times and hope that we can inspire even more people with our ElasticAI approach in the future!

We are pleased to announce that our external PhD student Sam Erbati successfully defended his dissertation on January 29th. Under the title "Towards the optimal orchestration of service function chains to enable ultra-reliable low latency communications in an NFV-enabled network", he made a significant contribution to the research field of "Network Function Virtualization".

His work focused on developing optimized orchestration methods for service function chains to enable ultra-reliable low latency communications in NFV-enabled networks.

On behalf of Professor Schiele and the entire department, we congratulate him on this milestone and wish him all the best for future challenges!

The RIWWER project is intended to benefit not only Duisburg's wastewater system, but also other municipalities throughout Germany.

The project findings are to be incorporated into the VDI expert recommendation 4900 on the topic of digitalisation of wastewater management. Expert recommendations are comparable to groundwork for VDI guidelines.

This enedeavour was launched on 6 December at the VDI premises in Düsseldorf. The project members and representatives of the water management sector met there to generalise the results for other wastewater systems.

This week, a delegation from our department visited the EAI MobiQuitous, which took place in Melbourne, Australia this year.

Chao Qian and Peter Zdankin (Distributed Systems) chaired the AutoQuitous workshop, which was dedicated to autonomous systems on both a technical and ethical level.

As part of this workshop, Tianheng Ling presented a paper co-authored with Qian entitled On-Device Soft Sensors: Real-Time Fluid Flow Estimation from Level Sensor Data.

That's a first: As a result of the long-standing cooperation between the Akian College of Science & Engineering of the American University of Armenia and the Faculty of Engineering, an Armenian-German research summit on the topic of "Intelligent and Data-Based Systems" is being held for the first time.

The summit will take place from 31 July to 2 August in Yerevan. Among the participants are members of the AUA, our department as well as researchers from the Universidad de Chile.

On 23 June, part of our department around the IoT Garage visited the FabLab Lothar of the Department of Manufacturing Engineering.
Along with a guided tour through the hall (!) and a presentation of the equipment, we had an exchange about how the day-to-day business and organizational matters are run.

Update: On 7 July we subsequently showcased the IoT Garage, including equipment, workstations, courses and projects, and of course many topics were discussed as well, ranging from 3D printing to publicity.

We thank the FabLab for visiting us and, not least, for the maany suggestions and impulses!

Today, our elastic-AI.runtime team gave a presentation in the IoT Lab about the progress made in developing the runtime over the last year, including a demonstration of the web-based user interface.

Afterwards, there was a lively discussion about the project and also future features.

Our student employees are thusly making an essential contribution to the elastic-AI ecosystem!

We are proud to announce that we won the Best Demo Award at the IEEE PerCom 2023 in Atlanta, Georgia!

Unfortunately, Chao Qian and Tianheng Ling could not be there, so our employees Lukas Einhaus and Leo Buron presented the demo. The award is therefore a true team achievement and we would also like to thank the two presenters!

At this year's PerCom in Atlanta, our department made several appearances:

On the first day, Prof. Schiele gave the keynote talk of the *LESS workshop, in which he presented the Elastic AI ecosystem to a broad audience.

On the same day, Tianheng Ling gave a remote presentation titled "On-device AI: Quantisation-aware Training of Transformers in Time-Series" as part of the PhD Forum.

Last but not least, on the second day of the conference, Lukas Einhaus and Leo Buron gave a practical demonstration of Elastic AI in action.

While the UDE was offline, our department attracted a new member: Dr.-Ing. Andreas Erbslöh.

He now adds the perspective of an electrial engineer to our team, i.e. a hardware-aware perspective.

His expertise is currently employed in the processing of neural signals in the Sp:AI:ke project.

Welcome to the team!

What are the implications of allowing critical decisions to be increasingly made by autonomous intelligent computer systems without any human involvement? Prof. Schiele is one of four experts who will discuss this matter online on March 24th as part of the panel at IEEE PerCom 2022, a world-leading conference for such systems.

Our application on the topic "Neural signal processing using artificial intelligence on an embedded platform" was accepted within the 21st call of the Mercator Research Center Ruhr (MERCUR)'s funding line MERCUR.KOOPERATION. Together with Karsten Seidl (UDE) and Christian Klaes (RUB), we are researching how to integrate AI into a medical implant.