Center for Artificial Intelligence and Robotics
CAIRO – Part of the Data Science Node in the AI District of Bavaria
In order to meet the paradigm shift towards a “digitized country”, the Free State of Bavaria announced the “Hightech Agenda Bayern” in its government declaration on 10 October 2019. In conjunction with the acceleration programme “Hightech Agenda Plus”, this means an investment boost of 3.5 billion euros – mainly profiting the Bavarian universities as well as universities of applied sciences.
One of the four key areas of this investment initiative is supporting Artificial Intelligence (AI) and Supertech, i. e. quantum technology, aerospace technology, and cleantech: The aim is to establish a Bavarian AI-network with the center located in Munich and various nodes and endpoints distributed all over the state. For THWS, this resulted in the foundation of the Competence Center Artificial Intelligence and Robotics (CAIRO). Together with the Center for Artificial Intelligence in Data Science (CAIDAS) at the Julius-Maximilians-Universität of Würzburg, CAIRO will be forming the data science node – whilst the Center Robotics (CERI) in Schweinfurt will be the end node.
CAIRO is closely linked to its THWS sister center CERI. Both centers are working towards the integration of AI into Robotics (CAIRO) as well as the determination of AI requirements within Robotics (CERI).
In order to realize these goals, the Free State of Bavaria has already established three professorships for the establishment of CAIRO. THWS has contributed an additional four research professorships and will be establishing further AI-professorships in various faculties at THWS.
For more details on CAIRO, please visit cairo.thws.de
Research focus on applied “strong AI”
The objective of the Center for Artificial Intelligence and Robotics (CAIRO) is to take up current topics in artificial intelligence and research them for the sake of society while facilitating effective knowledge transfer to the practical domain within various industry sectors. The focus is also on a more holistic understanding of AI – as opposed to isolated individual scenarios, which is the typical practice today. Ethical and legal aspects of AI will play an important role.
Starting from the fundamental areas of intelligence (perceiving, reasoning and acting) along with a concept-based representation of knowledge, it is possible to distinguish interlinked research fields at the Center for Artificial Intelligence and Robotics (CAIRO). These research fields reflect how CAIRO is organized: perception with conceptualization (perceiving), problem recognition and conceptual solving as well as realization of action concepts (acting) and conceptual learning (central reasoning).
Characterization of CAIRO's six functional fields of research
AI perception with conceptualization (AI perception)
Signals and (sensory) stimuli must be detected (using intelligent sensors) and generally converted with the aid of situational knowledge into a perception concept (concept of what is perceived), which can then be interpreted and processed by a central reasoning apparatus. Besides natural language processing and computer vision, models and theories must be applied in order to operationalize the conceptualization processes.
AI problem recognition (AI recognition)
One key reasoning ability involves recognition of the need to utilize an available perception concept (problem recognition). The utilization can take place through learning (inward utilization) or through derivation of an action concept (outward utilization).
AI solution development (AI solving)
A situation that was assessed as requiring change necessitates correction on the basis of an action. For this purpose, available knowledge must be used to define a target or solution concept, which must then be subsequently materialized through action.
AI realization of action concepts, motion (AI motion)
Necessary changes and solution concepts that relate to physical objects call for realization of a motion sequence involving an extremity or a peripheral apparatus (arms, legs, grippers, etc.) along with creation of a corresponding action concept.
AI realization of action concepts, communication (AI communication)
Necessary changes and solution concepts directed towards other intelligence media or which relate to them are to be realized as acts of speech and must be communicated in the form of spoken or text-based instructions, questions or statements. Necessary changes that are directed inwards or which relate to oneself are also acts of speech or communication that are addressed to one’s own reasoning apparatus (speaking with oneself).
AI conceptual learning (AI learning)
Through learning, a utilization concept is either supplied as a new insight to the reasoning apparatus or the knowledge base, or known concepts are extended or linked. Newly emerging or modified knowledge can in turn represent a need for utilization and trigger further reasoning processes.
Strong or weak AI?
In order to realize intelligent machines (software implementation) in these areas, it is possible to make use of weak as well as strong AI approaches, or a combination of the two.
Determination of which approach is best suited for a given area is an important responsibility for CAIRO.