Guest Editorial Special Issue on Cognitive Learning of Multiagent Systems
Yang Tang,Wei Lin,Chenguang Yang,Nicola Gatti,Gary G. Yen
DOI: https://doi.org/10.1109/tcds.2023.3325505
IF: 4.546
2024-02-07
IEEE Transactions on Cognitive and Developmental Systems
Abstract:The development and cognition of biological and intelligent individuals shed light on the development of cognitive, autonomous, and evolutionary robotics. Take the collective behavior of birds as an example, each individual effectively communicates information and learns from multiple neighbors, facilitating cooperative decision making among them. These interactions among individuals illuminate the growth and cognition of natural groups throughout the evolutionary process, and they can be effectively modeled as multiagent systems. Multiagent systems have the ability to solve problems that are difficult or impossible for an individual agent or a monolithic system to solve, which also improves the robustness and efficiency through collaborative learning. Multiagent learning is playing an increasingly important role in various fields, such as aerospace systems, intelligent transportation, smart grids, etc. With the environment growing increasingly intricate, characterized by factors, such as high dynamism and incomplete/imperfect observational data, the challenges associated with various tasks are escalating. These challenges encompass issues like information sharing, the definition of learning objectives, and grappling with the curse of dimensionality. Unfortunately, many of the existing methods are struggling to effectively address these multifaceted issues in the realm of cognitive intelligence. Furthermore, the field of cognitive learning in multiagent systems underscores the efficiency of distributed learning, demonstrating the capacity to acquire the skill of learning itself collectively. In light of this, multiagent learning, while holding substantial research significance, confronts a spectrum of learning problems that span from single to multiple agents, simplicity to complexity, low dimensionality to high dimensionality, and one domain to various other domains. Agents autonomously and rapidly make swarm intelligent decisions through cognitive learning overcoming the above challenges, which holds significant importance for the advancement of various practical fields.
robotics,computer science, artificial intelligence,neurosciences