PROGRAMME
AUTONOMY of INTELLIGENCE
and
BRAIN MACHINERY
AUTONOMY of INTELLIGENCE
Autonomous systems execute tasks by elaborating information on their environment captured by sensors.
The interactions of autonomous systems with humans characterize the level of autonomy:
- direct-interaction systems are almost completely controlled by an operator,
- operator-assisted systems require supervisory input from a human to choose and perform tasks,
- fully autonomous systems can operate without the assistance of an operator.
The subsystem of intelligent behaviors of an autonomous system consists in the execution of tasks satisfying goals in unpredictable environments and/or modifying their environment to satisfy goals.
Autonomy by Intelligence is the guiding principle to design and realize intelligent autonomous systems able to solve complex tasks with the minimum need of human intelligence.
Artificial Intelligence, Machine learning technologies and advanced sensors to collect information about the environment are the enabling technologies allowing the system to satisfy goals by interacting with their environment.
BRAIN MACHINERY
Brain machinery is the subsystem of an autonomous intelligent system allowing to support the subsystem of its intelligent behaviors through a perception, processing and action feedback loop:
- perception: collecting information from the environment of the systems through sensors and combining that data obtained also from the effect of the action of the system on the environment,
- processing: extracting relevant patterns of information by removing irrelevant data from perception to allow the system to take decisions,
- action: according to the taken decisions, acting by performing tasks needed to satisfy goals.
Brain machinery is realizable by a feedback loop between
- model based specifications of the environment, of the possible interactions of the environment and of the degree of satisfaction of such interactions with respect to goals,
- data-driven machine learning from the information on the interactions with the environment and from the degree of satisfaction of such interactions with respect to goals.
MORPHODYNAMICS of AUTONOMOUS INTELLIGENT SYSTEMS
FORMS and SUBSTRATUM
The design of Autonomous Intelligent Systems is inspiredby René Thom’s model theory based on his mathematical research on morphogenesis and by Alain Cardon’s studies on Artificial Intelligence.
The Alexandre Grothendieck’s sketch of“geometry of forms” is a source of inspiration to be further studied.
An Autonomous Intelligent System consists of two subsystems:
- the substratum consisting in the subsystem hybrid brain machinery and
- the space of forms consisting in the subsystem of intelligent behaviors.
Such approach can offer general principles to be developed for guiding the concrete study and realization of Autonomous Intelligent System in several applications domains.