Fraunhofer Institute for Digital Media Technology (Fraunhofer IDMT) is today launching its virtual exhibition, part of the IBC Showcase, which will take place 8-11 September, from 10am to 8pm CET daily.
At this year´s online IBC Showcase event, Fraunhofer IDMT will present its latest research activities and technologies that transform and improve media workflows.
Innovations from Fraunhofer IDMT’s extensive technology portfolio will be showcased including automatic metadata extraction and enrichment, audio and video matching, content verification, search and privacy-enhanced content recommendation.
These innovations combine audiovisual content analysis, machine learning and AI, security and privacy protection to form a suite of “Trusted AI Solutions” that enable media businesses to fully exploit the potential of their audiovisual content.
Meet Patrick Aichroth, Hanna Lukashevich and Uwe Kühhirt, Fraunhofer IDMT’s experts in machine learning, content analysis and trusted AI at the virtual showcase to learn about the latest innovations, current research agenda, and discuss hot topics:
The use of automatic content analysis to drive monetization of archives and the long tail, and to save costs via content tracking and de-duplication.
The use of audio forensics and matching to detect fakes and disinformation, and facilitate content verification and moderation.
Misconceptions, opportunities and risks around the deployment of Artificial Intelligence solutions.
“AI is not magic”, says Hanna Lukashevich, “many people see is as something that thinks for itself and is smart. “But they’re wrong: AI is mostly mathematics, and it can only do what it has been taught in advance. It is key to know beforehand what you want to use it for, and then select, adapt and evaluate AI considering your specific needs”.
“Similarly, concepts like bias and fairness can only be addressed if the target use case is clear, and if there is an agreed definition for what fairness means in that case”, comments Patrick Aichroth. “As for other technologies, it depends completely on how we use AI whether it is useful or problematic. Recommendation technologies, for instance, can reinforce human confirmation bias, or they can help us to lower confirmation bias, depending on how we design it.”
We invite you to set up a meeting with Hanna Lukashevich, Head of Semantic Music Technologies, Patrick Aichroth, Head of Media Distribution and Security and Uwe Kühhirt, Head of Metadata Department, to learn more about AI-driven solutions for the media sector, and get insights into our research activities to transform and improve media workflows.