iCog is an open source initiative started at IIT with the goal of advancing our knowledge of human cognition by designing, building, and sharing a common cognitive architecture for an embodied artificial system such as iCub.
The iCog cognitive architecture will provide a coherent baseline that will facilitate convergent, cumulative progress in the development of an operational model of cognition. In parallel iCog will facilitate experimental testing of core cognitive abilities and their dynamic and synergistic interplay as the robot interacts with its environment and other cognitive agents.
The intent is not necessarily to develop a single implementation of a cognitive architecture, but to establish a practical workshop environment where we can discuss and experiment with the core cognitive abilities, and discover how they are dynamically coordinated in the framework of the a complete cognitive system, encapsulated in the iCog cognitive architecture.
As such, the iCog initiative is planned to host and stimulate the convergence of various disciplines such as computer science, artificial intelligence, neuro & cognitive sciences, robotics, and social sciences. Each of these disciplines provides a different perspective and brings its own domain-specific approach, knowledge, data, and experimental methodologies to an integrated view of human-level cognition.
iCog will be implemented as a dynamically-adaptable software framework embedded in a physical body. As such, it will be a source of disruptive technologies supporting the development of personal robots and machines that “know what they are doing” and can interact with humans in a humane and predictive way, anticipating the needs of humans and the actions required to assist them.
More resources on cognition can be found on the website of IEEE Technical Committee for Cognitive Robotics.