Please use this identifier to cite or link to this item:
|Title:||A grand challenge for computational intelligence a micro-environment benchmark for adaptive autonomous intelligent agents||Authors:||Ho, S.-B.||Keywords:||adaptive systems
general principles of learning
|Issue Date:||2013||Citation:||Ho, S.-B. (2013). A grand challenge for computational intelligence a micro-environment benchmark for adaptive autonomous intelligent agents. Proceedings of the 2013 IEEE Symposium on Intelligent Agents, IA 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013 : 44-53. ScholarBank@NUS Repository. https://doi.org/10.1109/IA.2013.6595188||Abstract:||Being able to acquire knowledge and form concepts by observing, exploring, and interacting with the environment and then applying the knowledge thus gained for problem solving to satisfy its goals and needs is the hallmark of an adaptive autonomous intelligent agent. However, for an intelligent agent to be fully autonomous and adaptive, all aspects of intelligent processing from perception to action must be engaged and integrated. To build such an all-encompassing system is a formidable task. We propose that a good approach is to first identify the necessary intelligent computational structures and processes for dealing with a suitably designed micro-environment so that they are tractable. The challenge for computational intelligence is then to uncover general principles leading to general computational structures and processes that can deal with the micro-environment and that are also scalable to deal with more complex and real-world environments. Neuroscience research revealed that there are indeed such scalable general mechanisms in the brain and this is reviewed to provide inspirations for the building of artificial systems. A suitable micro-environment for this purpose must consist of a minimal set of features necessary to engage the various intelligent processes from that of the perceptual to that of the attentional, memory, affective, conceptual, planning, action, and learning. The micro-environment benchmark we propose here consists of an internal environment including the affective states of the intelligent agent as well as an external environment that is dynamic and in which activities of and interactions between objects can take place to engage the intelligent agent in all the intelligent processes described above. © 2013 IEEE.||Source Title:||Proceedings of the 2013 IEEE Symposium on Intelligent Agents, IA 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013||URI:||http://scholarbank.nus.edu.sg/handle/10635/111518||ISBN:||9781467358811||DOI:||10.1109/IA.2013.6595188|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
checked on May 15, 2021
checked on May 17, 2021
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.