E. Santos Jr. and D. O. Banks, "Acquiring Consistent Knowledge," Technical Report AFIT/EN/TR96-01, Air Force Institute of Technology, January 1996.  Home/Search   Document Details and Download   Summary   Related Articles

....on line. Still more complex hybrids can use nested agents within multi system func20 tion approximators and use the approximators to help higher level agents learn profiles and search databases and perhaps perform other agent tasks. The neural fuzzy agent needs to improve how it acquires knowledge [18, 52]. The agent should not ask the user too many questions. The agent needs to learn the user s profile fast enough before it tires the user. Efficient agents would make the user state rankings that are at most linear in the number of search objects or search object clusters. Our system asks the user ....

....system asks the user a large number of numerical questions even though the user may not want to give and perhaps cannot give precise numerical answers to these questions. Researchers have long searched for techniques that can lessen the number of numerical questions the system must ask the user [18, 52]. The bootstrap and other statistical methods [9] may offer more efficient ways for an adaptive agent to sample its user and its environment. Ordinal or chunking techniques [26, 31, 36] may also ease the burden of preference acquisition. But all such techniques tend to increase the complexity of ....

E. Santos Jr. and D. O. Banks, "Acquiring Consistent Knowledge," Technical Report AFIT/EN/TR96-01, Air Force Institute of Technology, January 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC