150 participants, speakers and exhibitors did not want to miss the top-class program of the Artificial Intelligence Forum despite Corona and followed the lectures live via the Internet on 14 May. Anyone who missed the virtual congress can now watch the expert presentations as recordings:

The presentations and videos of our virtual Artificial Intelligence Forum 2020 are also available for download for non-participants at a price of 349 euros. Please contact the event department of WEKA Fachmedien.

This year's Artificial Intelligence Forum had a lot to offer. In four keynotes, experts from business, science and industry spoke about the advantages and disadvantages of AI as well as the very specific challenges of dealing with artificial intelligence. But read for yourself:

"AI can constructively support people, but not creatively replace them"

The third AI Forum began with the presentation "AI + humans = unbeatable together!", given by Reinhard Karger, company spokesperson for the German Research Center for Artificial Intelligence (DFKI). He began by explaining that artificial intelligence (AI) refers to the digitalization of human knowledge skills, i.e. the attempt to have computers perform services for which humans require intelligence, with the aim of offering humans physical or cognitive assistance functions. The only question is how this can be achieved without the awareness that humans are ultimately irreplaceable. "AI can constructively support people, but not creatively replace them," emphasized Karger. Current AI applications enable many ways of supporting people, offering them new experiences or developing new products. But the variety of new tools does not yet lead to a general AI, a whole that is more than the sum of its parts. The focus is now on specific issues that SMEs can address, specific goals and ways in which AI can help to achieve them. Above all, Karger emphasized: "AI has a lot of potential, but common sense is unbeatable. Humans are masters at reducing complexity and then doing the right thing. We have a mastery that we masterfully underestimate."

"Ethics must not be instrumentalized to prevent AI."

Prof. Dr. Christoph Lütge from the Technical University of Munich then introduced his audience to ethics in artificial intelligence. After all, AI naturally has the potential to improve the lives of many people in the long term, for example in autonomous driving. But of course artificial intelligence can also have a negative impact on society. Clear rules are therefore necessary. These rules should not just be legal regulations; above all, ethical guidelines are needed to address widespread fears. After all, algorithms are only as good as the assumptions on which they are ultimately based. People need to develop trust in AI and if they don't trust the AI system, they won't use it, Lütge is convinced: "AI won't fly without ethics!" But he also warned: "Ethics must not be instrumentalized to prevent AI."

"It's not because we have a shiny new hammer that we should see every problem as a nail."

Dr. Steven Peters, who heads the "AI Research" team within Mercedes-Benz Group Research, spoke on the topic of "AI as a tool for engineers". His team is researching machine learning methods for use in future vehicle functions and as an assistance system for development engineers. In his presentation, Peters began by explaining the differences between classic algorithms and machine learning. Peters described how learning in the car works - from individual learning to learning from the fleet and empathic learning through to individual training and concluded by explaining the car manufacturer's AI principles to the audience: What does the responsible use of AI look like - "you always have to ask yourself the question, is there another approach that might be better suited to achieving your goal". What role do explainability, privacy protection, security and reliability play? Peters compared AI to a shiny new hammer: "It's not because we have a shiny new hammer that we should see every problem as a nail."

Domain-specific architectures offer adaptability

The third Artificial Intelligence Forum closed with a keynote from US chip manufacturer Xilinx, which makes extensive use of programmable logic to process AI algorithms. Ramine Roane, Vice President Software and AI Product Marketing, gave a presentation on "Challenges in AI production: adaptability from sensor to brain to action." Roane sees two major challenges: One is an order of magnitude increase in computing power, while keeping price, performance, latency and form factor intact. Secondly, an appropriate architecture is needed to meet the high demands of complex algorithms. In his opinion, this requires an adaptable domain-specific architecture that must be implemented from the sensor to the brain to the actions.

After the keynotes, the forum split into three sessions: Artificial Intelligence in Embedded Applications, AI in the Factory and Artificial Intelligence in Vehicles. You can find the program on the website of the Artificial Intelligence Forum.

___________________________________________________________________

The fourth Artificial Intelligence Forum is expected to take place again as a face-to-face event on 21.04.2021 in Stuttgart.

___________________________________________________________________

The presentations and videos of our virtual Forum AI 2020 are also available for download for non-participants at a price of 349 euros. Please contact the event department of WEKA Fachmedien.

___________________________________________________________________