Publication:
Inferring human values for safe AGI design

Simple item page
dc.contributor.authorSezener, Can Eren
dc.contributor.ozugradstudentSezener, Can Eren
dc.date.accessioned2016-06-30T12:33:30Z
dc.date.available2016-06-30T12:33:30Z
dc.date.issued2015
dc.description.abstractAligning goals of superintelligent machines with human values is one of the ways to pursue safety in AGI systems. To achieve this, it is first necessary to learn what human values are. However, human values are incredibly complex and cannot easily be formalized by hand. In this work, we propose a general framework to estimate the values of a human given its behavior.
dc.identifier.doi10.1007/978-3-319-21365-1_16
dc.identifier.endpage155
dc.identifier.isbn978-3-319-21365-1
dc.identifier.scopus2-s2.0-84952760405
dc.identifier.startpage152
dc.identifier.urihttp://hdl.handle.net/10679/4184
dc.identifier.urihttps://doi.org/10.1007/978-3-319-21365-1_16
dc.identifier.wos000363479400016
dc.language.isoengen_US
dc.peerreviewedyes
dc.publicationstatuspublisheden_US
dc.publisherSpringer International Publishing
dc.relation.ispartofLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
dc.relation.publicationcategoryInternational
dc.rightsrestrictedAccess
dc.subject.keywordsValue learning
dc.subject.keywordsInverse reinforcement learning
dc.subject.keywordsFriendly AI
dc.subject.keywordsSafe AGI
dc.titleInferring human values for safe AGI designen_US
dc.typebookParten_US
dc.type.subtypeBook chapter
dspace.entity.typePublication

Files

Collections

Computer Science