Prof. Dr. Marco Huber (Universität Stuttgart /IPA Fraunhofer)
Cognitive Production Systems – Machine Learning in Industrial Manufacturing

Machine learning (ML) methods have recently led to enormous progress in the field of artificial intelligence. It allows the automatic recognition and exploitation of correlations and patterns in complex data. The application of ML is particularly useful in applications where cause-effect relationships are very difficult or impossible to describe analytically using mathematical methods, but instead extensive data is available. This situation is encountered in many places in industrial manufacturing. Production facilities are continuously monitored by various sensors so that ML processes can be triggered, action plans can be generated and then executed, resulting in continuous optimizations of production processes.

In this talk, first a brief introduction to the topics of artificial intelligence and ML is given, together with highlighting the benefits and limitations. This will be followed by an introduction of basic ML principles. This is combined with providing insights to a large number of real-world use cases solved at Fraunhofer IPA together with different manufacturing companies.

Dr. Lars Mündermann (Karl Storz SE & Co. KG)
AI in medicine – overrated or groundbreaking?

Developments in the last decade have shown that technology is the most disruptive factor in healthcare. This trend is expected to continue in the upcoming decades as Health & Care is one of the sectors with highest levels of investment in new technologies, treatment options, and drugs.

Medical treatment is expected to be supported by a range of diagnostic tools, and real-life data on treatment success rates may influence the outcome for patients. Patients may no longer rely on a single and local physician but may access platforms whenever they have a specific medical need. Health might no longer be defined by the absence of a disease but rather be seen in the more holistic context of a person's wellbeing embracing the concept of P4 medicine (predictive, preventive, personalized and participatory).

This keynote will illustrate what health and medicine might look like in 2050, introduce research activities in the fields of Cognitive Surgery and Surgical Data Science, discuss potential for AI applications in areas such as digital monitoring, prevention and AI-assisted diagnostics, and allude to prerequisites and restrictions for (successfully) introducing AI in medicine.

Bogdan Penkovsky (Alysophil SAS)
Towards autonomous API synthesis with deep reinforcement learning

Reinforcement learning (RL) is a general learning and decision making paradigm based on interaction with the environment. Despite being challenging in practice, implementing RL for real world tasks could improve the safety and efficiency of autonomous processes. In this work we apply RL for an autonomous molecule synthesis with continuous flow chemistry. In our preliminary experiments, we demonstrate that our agent trained by deep reinforcement learning is capable of responding to real-time challenges, such as changes in the environment, sensor noise, and perturbations in order to ensure the optimal chemical synthesis conditions. The ultimate goal of this work is to conceive an autonomous chemical production unit for active pharmaceutical ingredients (API).