The Multipurpose Enhanced Cognitive Architecture(MECA)is a cognitive framework designed to model complex,human-like processes across multiple domains.Originally focusing on implementing a Dual Process Theory approach ...The Multipurpose Enhanced Cognitive Architecture(MECA)is a cognitive framework designed to model complex,human-like processes across multiple domains.Originally focusing on implementing a Dual Process Theory approach and integrating a machine consciousness mechanism based on Global Workspace Theory,MECA has been updated to integrate a dual-layer subsumption mechanism,enabling both reactive and deliberative behaviors,dynamic goal setting and a visual-spatial memory subsystem,enhancing MECA’s capacity for real-world interaction and adaptive behavior.Also,with the introduction of the new computational ideas’knowledge representation scheme,MECA proposes to organize knowledge dynamically to handle context-sensitive reasoning and flexible categorization.MECA’s implementation relies on the Cognitive Systems Toolkit(CST),facilitating its integration with cutting-edge technologies.MECA and CST are being continuously developed and updated,aligned,and open to incorporate the latest AI artifacts and methodologies.This approach ensures the delivery of organized,monitorable,auditable,and controllable AI solutions,significantly reducing reliance on“black box”cognitive processes while enhancing transparency and accountability in AI-driven systems.These updates reinforce MECA’s potential as a robust architecture for developing autonomous,adaptable,and context-aware AI systems capable of real-world interaction and adaptive learning.展开更多
基金Supported by the Sao Paulo Research Foundation(FAPESP),CPE SMARTNESS(2021/00199-8)and CEPID/BRAINN(2013/07559-3).
文摘The Multipurpose Enhanced Cognitive Architecture(MECA)is a cognitive framework designed to model complex,human-like processes across multiple domains.Originally focusing on implementing a Dual Process Theory approach and integrating a machine consciousness mechanism based on Global Workspace Theory,MECA has been updated to integrate a dual-layer subsumption mechanism,enabling both reactive and deliberative behaviors,dynamic goal setting and a visual-spatial memory subsystem,enhancing MECA’s capacity for real-world interaction and adaptive behavior.Also,with the introduction of the new computational ideas’knowledge representation scheme,MECA proposes to organize knowledge dynamically to handle context-sensitive reasoning and flexible categorization.MECA’s implementation relies on the Cognitive Systems Toolkit(CST),facilitating its integration with cutting-edge technologies.MECA and CST are being continuously developed and updated,aligned,and open to incorporate the latest AI artifacts and methodologies.This approach ensures the delivery of organized,monitorable,auditable,and controllable AI solutions,significantly reducing reliance on“black box”cognitive processes while enhancing transparency and accountability in AI-driven systems.These updates reinforce MECA’s potential as a robust architecture for developing autonomous,adaptable,and context-aware AI systems capable of real-world interaction and adaptive learning.
