As part of AI Power, we put you in the spotlight CognoRobotics which is a general purpose Robotics Cognitive Developmental Architecture. In other words, it’s the development of autonomous robots. For the development of Artificial General Intelligence and its application to the field of autonomous robotics, the development of a Mathematical and Computational Framework of Autonomous Cognitive Development is essential.
A promising project for an innovative future
CognoRobotics integrates differents AI paradigms :
- Integrative Computational Cognitive Architecture: the proposed architecture must integrate all cognitive components: perceptual, motor, decision making, planning, learning, development, etc.
- “Ground-up” Cognitive Development: development that starts from the most basic and sensorimotor experiences.
- Embodied Agent: the body is guided by the mind to interact actively and purposefully with its environment, at the same time, the body guides the mind thanks to self-perception.
- Incremental and Modular Architecture: to develop efficient computing and learning, it is essential to reuse the components from one experience to another.
- Hierarchical and Multilevel Architecture: It must work on multiple levels that interact and are integrated: sensorimotor, hierarchical planning, abstract concepts, abstract reasoning, etc.
- Anticipatory at all levels: that is, all processes, from the most basic sensorimotor functions, to the most abstract of processes, must include the ability to anticipate (predict) the effect that the execution of that process will have, both, in the environment and in the agent.
- Mathematically solid framework: only in this way can demonstrations of convergence, consistency, completeness, optimality, etc. be carried out. It is based on the mathematical formalization of Probabilistic Graphs, Markov Decision Processes, Bayesian Structure Learning and Causal Learning.
In collaboration with various research groups around the world, we have already worked previously on some of the elements necessary to build this Integrated Architecture for Cognitive Development. We have already begun developing a very preliminary deterministic version of SBL at the University of Southern California. In Spain, at the Autonomous University of Madrid, we developed a Behavior-based robot. We have already developed several Innovation projects such as, for example, DriveSense, Hyperios, ACT, that have allowed the development of different robots with different architectures and with different specific objectives. We have extensive experience in Robotics for Industry and Defense customers. We are also currently extending to other sectors where there are physical sensors, such as in Medicine.
The objectives of the project
The main objective of this project consists of the research and development of a Mathematical and Computational Framework for Cognitive Development in Autonomous Robotics. In this regard, the two main objectives of the project are, on the one hand, the R&D of a Cognitive Computational Architecture that allows the cognitive development of Robots and Autonomous Agents. On the other hand, the R&D of a sufficiently broad Mathematical Framework that serves to explain the behavior in biological agents, as well as to design consistent and optimal artificial agents.
Article writed with Fernando Corbacho of AI Power team
About AI Power
This is an innovative approach to enable companies to capitalize on this technology. This support, led by Audensiel’s AI expert team, will help develop this evolution and adapt it to the business uses of companies.
Comments