國立虎尾科技大學 |

Neuro-Symbolic AI : = A Probabilistic Perspective.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Neuro-Symbolic AI :/
Reminder of title:	A Probabilistic Perspective.
Author:	Ahmed, Kareem Ahmed Yousry Abdelraof.
Description:	1 online resource (300 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-12, Section: A.
Contained By:	Dissertations Abstracts International85-12A.
Subject:	Computer science. -
Online resource:	click for full text (PQDT)
ISBN:	9798382832951

Neuro-Symbolic AI : = A Probabilistic Perspective.
Ahmed, Kareem Ahmed Yousry Abdelraof.

Neuro-Symbolic AI :A Probabilistic Perspective. - 1 online resource (300 pages)

Source: Dissertations Abstracts International, Volume: 85-12, Section: A.

Thesis (Ph.D.)--University of California, Los Angeles, 2024.

Includes bibliographical references

The last decade has witnessed an explosion of interest in Artificial Intelligence, not only among researchers, but also in the public eye. This has led to machine learning (ML) systems being increasingly deployed in domains with high personal and societal impact. Consequently, there's a growing need to reason about the behavior of these machine learning systems, with the ultimate goal of mitigating unwanted behaviors, ensuring reliable outputs, and fostering trustworthy interactions with the end user. This dissertation lays the foundations for reasoning about the behavior of such systems, especially when we have access to domain knowledge-sets of rules, or constraints, that characterize the set of valid predictions given the problem at hand. In particular, a core contribution of this dissertation is viewing ML systems as inducing probability distributions over output spaces. To reason about the behavior of these systems, one must then reason about the behavior of the underlying probability distributions. Building upon that perspective, the dissertation begins by developing methods that minimize the probability of ML systems producing invalid outputs, even when the constraints and distributions are theoretically intractable. It then goes further, developing methods that provide guarantees on the outputs of these ML systems, developing probabilistically-sound approaches to gradient estimation when the constraints are embedded within the architecture. The developed methods make use of tractable circuits whose structure differs from that of typical ML architectures. This dissertation, therefore, develops frameworks for expressing constraints as Python functions that can then be efficiently computed on GPUs. Lastly, this dissertation showcases the scalability and efficacy of the developed methods on real-word applications.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798382832951Subjects--Topical Terms:

573171
Computer science.
Subjects--Index Terms:

Machine learningIndex Terms--Genre/Form:

554714
Electronic books.

Neuro-Symbolic AI : = A Probabilistic Perspective.
LDR:03226ntm a22003977 4500 001 1146375
005 20240812064421.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798382832951
035 $a (MiAaPQ)AAI31331716
035 $a AAI31331716
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Ahmed, Kareem Ahmed Yousry Abdelraof. $3 1471758
245 1 0 $a Neuro-Symbolic AI : $b A Probabilistic Perspective.
264 0 $c 2024
300 $a 1 online resource (300 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: Dissertations Abstracts International, Volume: 85-12, Section: A.
500 $a Advisor: Van den Broeck, Guy;Chang, Kai-Wei.
502 $a Thesis (Ph.D.)--University of California, Los Angeles, 2024.
504 $a Includes bibliographical references
520 $a The last decade has witnessed an explosion of interest in Artificial Intelligence, not only among researchers, but also in the public eye. This has led to machine learning (ML) systems being increasingly deployed in domains with high personal and societal impact. Consequently, there's a growing need to reason about the behavior of these machine learning systems, with the ultimate goal of mitigating unwanted behaviors, ensuring reliable outputs, and fostering trustworthy interactions with the end user. This dissertation lays the foundations for reasoning about the behavior of such systems, especially when we have access to domain knowledge-sets of rules, or constraints, that characterize the set of valid predictions given the problem at hand. In particular, a core contribution of this dissertation is viewing ML systems as inducing probability distributions over output spaces. To reason about the behavior of these systems, one must then reason about the behavior of the underlying probability distributions. Building upon that perspective, the dissertation begins by developing methods that minimize the probability of ML systems producing invalid outputs, even when the constraints and distributions are theoretically intractable. It then goes further, developing methods that provide guarantees on the outputs of these ML systems, developing probabilistically-sound approaches to gradient estimation when the constraints are embedded within the architecture. The developed methods make use of tractable circuits whose structure differs from that of typical ML architectures. This dissertation, therefore, develops frameworks for expressing constraints as Python functions that can then be efficiently computed on GPUs. Lastly, this dissertation showcases the scalability and efficacy of the developed methods on real-word applications.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 573171
650 4 $a Linguistics. $3 557829
650 4 $a Information technology. $3 559429
653 $a Machine learning
653 $a Tractable circuits
653 $a Probability distributions
653 $a ML systems
655 7 $a Electronic books. $2 local $3 554714
690 $a 0800
690 $a 0984
690 $a 0489
690 $a 0290
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a University of California, Los Angeles. $b Computer Science 0201. $3 1182286
773 0 $t Dissertations Abstracts International $g 85-12A.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31331716 $z click for full text (PQDT)