The brain as a thinking machine
From biomass brain to electronic brain
Human consciousness and intelligence go well beyond mere information processing. Despite intensive research efforts, we still do not know what is actually involved in human reasoning and consciousness. Of course, the biochemical processes in the brain have been precisely analysed and studied, and scientists have worked for centuries to understand how humans act and think in the context of perception and consciousness. However, we are still unaware of how to artificially replicate these capabilities.
The fact that the natural medium of our intelligence, i.e. the “brain”, is a material object subject to the laws of physics leads to the logical assumption that it should be possible to fully replicate what our brain does by means of computers. This has already been demonstrated in the domain of computing: Artificial calculating machines are faster and more efficient and accurate than the human brain.
Replicating a “thinking machine”
Even in the age of computers, digitalization and (weak) artificial intelligence, a thinking machine based on the human model represents a special challenge. An approach involving the creation of artificial neural networks (ANNs) based on the model of our brain has led to enormous achievements in AI – surpassing human abilities in the areas of perception and pattern recognition (e.g. effortless recognition of image phenomena). However, some further characteristics of natural intelligence (see figure below) have not yet been successfully replicated on an artificial basis.
The functions of human intelligence range from perception by means of various sensory organs, through problem recognition, learning and solution development, and extending to a complete action concept capable of changing the perceived environment as desired by the intelligence medium when realized through motion or communication. Until now, self-determined collection, handling and usage of information and knowledge have also been reserved for natural intelligence media.
Realizing intelligence with strong AI
Strong AI with the eight fields of intelligence can be realized through machine conceptualization (creation and verification of concepts by machines), thus allowing realization of the main areas of intelligence: perceiving, reasoning and acting.
It is our vision to create an artificial thinking machine that possesses the capabilities of the human brain.
Prof. Robert Grebner, President of the University of Applied Sciences Würzburg-Schweinfurt
From mechanical machine to thinking machine
Based on the chronology of technologies and machines created by humankind over the different epochs, ranging from early mechanical machines to calculating machines, we can extrapolate the logical future development of machines: On the basis of digitalization, strong artificial intelligence will be manifested in a thinking machine. The era of intellectualization of machines is beginning. The revolutionary development step to follow will entail the emancipation of machines, i.e. machines will possess autonomous intelligence and become independent of human input and programming.
Although it is very difficult to imagine at present, the ultimate destination represents the humanization of machines, i.e. strong AI within humanoid robots: Humans and machines will thus exhibit equivalent levels of intelligence and capacity for action.
The epochs of machine systems
Began in 1769 with mechanical machines
(especially the steam engine invented by James Watt)
Foundation for the next epoch based on heavy turbines and generators
Began in 1866 with the dynamo generator by Werner von Siemens
Foundation for the next epoch based on semiconductors
Began in 1948 with the transistor
Foundation for the next epoch based on integrated circuits (ICs)
Began in 1937 with the Z1 by Konrad Zuse
Foundation for the next epoch based on networking, massive sensor deployment and thus the Internet of things as well as big data
Began in 2002 (assuming this was the first year in which more data was stored digitally than with analog technology)
Foundation for the next epoch based on artificial intelligence
Thinking machines with human-comparable intelligence (strong AI)
Foundation for the next epoch based on legal certainty and acceptance of intellectualization
Liberation of machines from human data input
Foundation for the next epoch based on human-like robotic machines
Strong artificial intelligence in machines that resemble humans and can move like humans
CAIRO: Incubator for applied strong AI
FHWS is concentrating its AI research activities at the Center for Artificial Intelligence and Robotics (CAIRO). The focus is on a more holistic understanding of AI – as opposed to isolated individual scenarios, which is still the typical practice today. In particular, techniques are being developed that involve a combination of strong AI and applied AI. Strong AI means that software that is developed can think and act independently in a problem-oriented manner – similar to (or just like) the human brain. Therefore, the relevant ethical and legal aspects of AI also play an important role at CAIRO.
Objectives and vision
The objective of CAIRO is
- to carry out interdependent research in the eight connected fields (see figure above) related to human cognitive intelligence,
- to integrate these fields into an application system, and thus
- to realize a comprehensive approach for artificial general intelligence (AGI) or strong AI.
Research results from the Bavarian AI nodes that have been established through the Hightech Agenda Bavaria should also be applied. In particular, results and findings produced at the Center Robotics and Intelligence in Schweinfurt (see CERI website and Robotics website) should also be integrated.
From CAIRO, a system should emerge offering integrated solutions for the eight fields of intelligence that is capable of learning elements such as
- how to deal with language,
- with physical movement and
- with its own self
in the context of social interaction and the physical environment, and thus establish a system of values (impression) along the lines of what our human brain is capable of.