Visions for AI research
Researchers of the RAI Cluster pool ideas
2024/06/18
Scientists from the “Reasonable Artificial Intelligence” (RAI) project have exchanged visions and methods for researching artificial intelligence (AI) as part of a series of retreats. The findings will feed directly into the application process for a Cluster of Excellence in the prestigious Excellence Strategy of the German federal and state governments.
What will future AI research look like? What methods will be used? The researchers discussed answers to these and other questions at the retreats of the RAI cluster project. The aim of RAI is to develop the next generation of AI, a “Reasonable Artificial Intelligence”. The research is divided into four areas: Systemic AI, Observational AI, Active AI and Challenging AI.
The researchers have used the previous dates to intensively discuss which methods can be used to realise their visions in the four areas and which supporting structures are required for this. TU President Professor Tanja Brühl and the Vice President for Research, Professor Matthias Oechsner, also took part in the opening discussion. The event was hosted by hessian.AI, the Hessian Centre for Artificial Intelligence.
Research areas RAI
RAI focuses on four research areas: Systemic AI works on software and system methods that enable efficient training of modular, decentralised RAI systems, which in particular combine learning with “thinking”, and support their integration into existing systems. The area Observational AI rethinks contextual learning and brings together different AI regimes to inject common-sense knowledge. The area Active AI focuses on continual and adaptive lifelong learning with active exploration. Finally, the area Challenging AI develops challenges and benchmarks on AI and RAI systems to test how much they “understand” the world in the human sense of “understanding”.
About RAI
Over the past decade, deep learning (DL) has enabled significant advances in artificial intelligence (AI), yet current AI systems have weaknesses, including a lack of logical reasoning, difficulties in dealing with new situations and the need for continuous adaptation. Last but not least, current AI systems require extensive resources. RAI aims to develop the next generation of AI: AI systems that learn with a “reasonable” amount of resources based on “reasonable” data quality and “reasonable” data protection. These are equipped with “common sense” and the ability to deal with new situations and contexts, and are based on sensible training paradigms that enable continuous improvement, interaction and adaptation.
Hessian.AI/cst
