Artificial intelligence is now an integral part of many technical systems. Artificial intelligence (AI), machine learning, deep learning and neural networks are opening up promising paths for further development in numerous areas of application. They are key technologies that enable systems to react autonomously and make decisions independently based on external influences.

The program

The Artificial Intelligence Forum, which the trade media Computer&AUTOMATION, Elektronik and Elektronik automotive are once again organizing virtually on 21 April 2021, will shed light on the rapid developments in hardware and software.

The keynote

This year, Prof. Dr. Marco Huber from the Fraunhofer Institute for Manufacturing Engineering and Automation IPA will open the virtual expert meeting with his keynote speech "Reliable AI - Safeguarding artificial neural networks". In it, he will address use cases for AI in which it is not accuracy but security and trust in the algorithms that are of enormous importance. In such applications, it is important that critical algorithmic decisions are safeguarded. In his lecture, Prof. Dr. Huber will look at the aspects of reliable and trustworthy AI. In particular, the aspect of safeguarding machine-learned models with a focus on deep artificial neural networks will be highlighted.

Artificial intelligence in the factory

In the "AI in the factory" session, we welcome Dr. Christoph Eichhorn from Sick, among others. In his presentation "Deep Learning: From hype to real customer benefits", he will explain how his company has used deep learning to improve its own applications in logistics automation and how this knowledge has been incorporated into product development. Eichhorn uses examples to show that deep learning is not only suitable for mastering even complex tasks, but that intuitive workflows also greatly simplify and accelerate solution development. The new deep learning tools can be used without detailed prior knowledge and are therefore accessible to a wider range of users.

3D environment detection for robots

In her presentation "3D environment detection for robots", Dr. Birgit Hasenberger from BECOM Systems will present a system for object classification and person recognition in the environment of an autonomous machine that is based on depth data from 3D time-of-flight sensors and artificial intelligence. Autonomous machines in particular are confronted with a large number of new requirements, such as complex navigation and handling tasks, unstructured environments and human-machine cooperation. The system presented consists of customized 3D sensor modules, an NVIDIA-based computing unit and deep learning software. As part of a feasibility study on person and object recognition, Dr. Hasenberger will demonstrate the complete workflow from the acquisition of training data to the deployment of such solutions with a deep learning platform.

Safety and AI - how does it work?

In his presentation "AI Classification Scheme for Safety", Holger Laible from Siemens discusses the field of tension surrounding the use of artificial intelligence (AI) in the context of safety applications. In addition to the risk assessment of application examples, an AI classification scheme will be presented, as is already being discussed in standardization work (ISO SC42). Approaches are set out as to which aspects should be considered for the further development of recommendations and normative requirements and current focal points from AI standardization are presented.

Avoid mistakes with AI

Rejects and rework in production and assembly cause costs and reduce competitiveness. Products and manufacturing processes are becoming increasingly complex and the number of possible influencing variables and dependencies is growing. As a result, rejects, rework and warranty risks are constantly increasing. This is where the IntelPromo - Intelligent Production and Assembly research project comes in, which the Institute for Machine Tools and Industrial Management (iwb) at the Technical University of Munich (TUM) is conducting in collaboration with Continental and mts Consulting und Engineering GmbH until December 2020. Automated process monitoring parallel to operation and the AI software Analyser are used to identify potential errors during the assembly process in real time and prevent them by intervening to control them. Frank Thurner from Contech Software & Engineering will present the project and the results in his presentation "Putting an end to rejects and rework - avoiding errors automatically with AI" .

With cloud and edge solutions for efficient production

"Intelligent edge solutions for more productivity and efficiency in the predictive maintenance of machines and systems" is the title of the presentation by Rudolf Krumenacker, Emerson Automation Solutions, ICC Intelligent Platforms. Due to increasing digitalization, every device, every machine, every system, every building will soon be networked via edge devices or connected to the cloud and vast amounts of data will be available for intelligent real-time evaluation. The presentation will show how this amount of data can be used for predictive maintenance and servicing of machines, plants and systems with intelligent edge devices and cloud connections, and how this is already being realized today. The presentation will also address the challenges, significance and scope of cyber security in connection with edge and cloud computing. The presentation will be rounded off with an outlook on possible future scenarios.

All information and the full program of the Artificial Intelligence Forum can be found at www.forum-ki.eu