We shape the future of the Internet of Things

IoT meets AI 2019 – Challenges around Industrial AI by Denis Krompass (Siemens)

At IoT meets AI 2019, Denis Krompass from Siemens talks about “Challenges around Industrial AI.”

Today’s most successful applications of AI are enabled by massive Machine Learning systems, especially from the domain of Deep Learning that besides scientific progress rely on a combination of two ingredients, computational power and huge amounts of labeled data. Not only do these demands have an increasing impact on our environment in terms of C02 emissions, the required labeled data amounts are a major bottleneck which limits the applicability of Deep Learning based AI to a broader market. Especially for the industrial domain where the data and problems are highly domain specific, this poses a major challenge, since the generated data can only be labeled by a few domain experts. This fact is opposed to the common domain, where basically anybody can label the data, e.g. decide if an image shows a cat or a dog. For this reason, one important aspect of industrial AI is to research methodologies that can deal with these data scarcity constraints.

Denis Krompaß, Ph.D., Siemens AG, Germany (Expert AI)

Denis Krompaß is Senior Key Expert for Deep Learning at Siemens Corporate Technology and the research lead of the Siemens AI Lab. During his time at Siemens he has researched and developed solutions for a wide set of applications that span, Factory automation, Energy Production, Healthcare and more.  His major research interest lies in leveraging scarce data scenarios with Transfer Learning. In the Siemens AI Lab he is currently ramping up the Siemens AI Lab Residency Program that will further expose challenges of Industrial AI to public research. In 2015 he got his PhD in Representation Learning on Knowledge Graphs supervised by Prof. Dr. Tresp at the LMU Munich where he is teaching Deep Learning since 2018.

Olivia Pahl

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

WP2Social Auto Publish Powered By : XYZScripts.com