國立虎尾科技大學 |

Computational level accounts of belief formation and revision in humans.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Computational level accounts of belief formation and revision in humans./
作者:	McDonnell, John V.
面頁冊數:	1 online resource (181 pages)
附註:	Source: Dissertation Abstracts International, Volume: 75-07(E), Section: B.
標題:	Experimental psychology. -
電子資源:	click for full text (PQDT)
ISBN:	9781303805837

Computational level accounts of belief formation and revision in humans.
McDonnell, John V.

Computational level accounts of belief formation and revision in humans. - 1 online resource (181 pages)

Source: Dissertation Abstracts International, Volume: 75-07(E), Section: B.

Thesis (Ph.D.)--New York University, 2014.

Includes bibliographical references

The work presented here reflects an effort to understand how humans form and revise beliefs, using probabilistic models as a tool for building theories of belief revision. Cognitive models were developed in two domains: semi-supervised category learning and causal reasoning with inconsistent information. These models provide a computational-level account of human learning and reasoning. At their core, the models assert that human reasoning and learning processes may be represented as parameterized probability distributions that are updated in light of experience.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9781303805837Subjects--Topical Terms:

1180476
Experimental psychology.
Index Terms--Genre/Form:

554714
Electronic books.

Computational level accounts of belief formation and revision in humans.
LDR:03493ntm a2200373K 4500 001 913730
005 20180622095236.5
006 m o u
007 cr mn||||a|a||
008 190606s2014 xx obm 000 0 eng d
020 $a 9781303805837
035 $a (MiAaPQ)AAI3614889
035 $a (MiAaPQ)nyu:11720
035 $a AAI3614889
040 $a MiAaPQ $b eng $c MiAaPQ
100 1 $a McDonnell, John V. $3 1186685
245 1 0 $a Computational level accounts of belief formation and revision in humans.
264 0 $c 2014
300 $a 1 online resource (181 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertation Abstracts International, Volume: 75-07(E), Section: B.
500 $a Adviser: Todd M. Gureckis.
502 $a Thesis (Ph.D.)--New York University, 2014.
504 $a Includes bibliographical references
520 $a The work presented here reflects an effort to understand how humans form and revise beliefs, using probabilistic models as a tool for building theories of belief revision. Cognitive models were developed in two domains: semi-supervised category learning and causal reasoning with inconsistent information. These models provide a computational-level account of human learning and reasoning. At their core, the models assert that human reasoning and learning processes may be represented as parameterized probability distributions that are updated in light of experience.
520 $a In the case of semi-supervised learning, many existing category learning models imply predictions about how the presence of unlabeled items will affect the generalization of category labels. Existing work has led to an unclear picture of how people integrate across these two types of learning. Here, I present a modified version of Anderson's (1991) rational model of categorization that is able to capture these effects. The resulting model predicted effects on participants' generalization strategies in three experiments. Empirical evidence was found in favor of these predictions in each experiment.
520 $a In the case of causal reasoning, a model was developed with a focus on coherence in causal reasoning. When learning about a domain, participants may learn different facts about the world which have incoherent implications. However, if we assume that participants' beliefs are encoded as probability distributions, their judgments may reflect a selection process in which incoherent outcomes are excluded. Applying this insight to the example of causal reasoning, a paradigm was developed in which participants were provided with incoherent facts. As predicted, their judgments appeared to reflect adjustments in the direction of coherent beliefs.
520 $a The data for these projects was collected on participants recruited on the Internet using Amazon's Mechanical Turk service and administered via their own web browsers. To validate this online methodology as a tool for studying human cognition, work was also done to replicate classic studies from the categorization literature.
520 $a Taken together, these studies offer new insight into our understanding of how people adapt their behavior in light of experience.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Experimental psychology. $3 1180476
650 4 $a Cognitive psychology. $3 556029
655 7 $a Electronic books. $2 local $3 554714
690 $a 0623
690 $a 0633
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a New York University. $b Psychology. $3 1183858
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3614889 $z click for full text (PQDT)