國立虎尾科技大學 |

Advanced Machine Learning Methods : = Semi-Supervised Learning, Multi-Category Classification, and Energy-Based Models.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Advanced Machine Learning Methods :/
Reminder of title:	Semi-Supervised Learning, Multi-Category Classification, and Energy-Based Models.
Author:	Zhang, Xinwei.
Description:	1 online resource (207 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-05, Section: B.
Contained By:	Dissertations Abstracts International85-05B.
Subject:	Statistics. -
Online resource:	click for full text (PQDT)
ISBN:	9798380851749

Advanced Machine Learning Methods : = Semi-Supervised Learning, Multi-Category Classification, and Energy-Based Models.
Zhang, Xinwei.

Advanced Machine Learning Methods :Semi-Supervised Learning, Multi-Category Classification, and Energy-Based Models. - 1 online resource (207 pages)

Source: Dissertations Abstracts International, Volume: 85-05, Section: B.

Thesis (Ph.D.)--Rutgers The State University of New Jersey, School of Graduate Studies, 2023.

Includes bibliographical references

This dissertation explores diverse aspects of machine learning with a particular focus on semi-supervised learning, multi-category classification, and energy-based models. The work begins by extending the statistical equivalence between logistic regression and exponential tilt modeling, leading to the development of a novel semi-supervised logistic learning method. Enhanced prediction accuracy is achieved by leveraging both labeled and unlabeled data. Next, we delve into the construction of loss functions and the establishment of corresponding regret bounds for multi-category classification. Here, new general representations of losses are derived, leading to the discovery of new hinge-like convex losses and multi-class proper scoring rules. In the final part, we introduce diffusion data and propose to learn a joint energy-based model, through persistent training, which simultaneously achieves long-run stability, post-training image generation, and superior out-of-distribution detection. Overall, this dissertation provides valuable insights and advancements in the field of machine learning, revealing connections between different methods and contributing to the improvement of existing methods and theories.

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

Mode of access: World Wide Web

ISBN: 9798380851749Subjects--Topical Terms:

556824
Statistics.
Subjects--Index Terms:

Energy-based modelsIndex Terms--Genre/Form:

554714
Electronic books.

Advanced Machine Learning Methods : = Semi-Supervised Learning, Multi-Category Classification, and Energy-Based Models.
LDR:02677ntm a22003857 4500 001 1143858
005 20240517105011.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798380851749
035 $a (MiAaPQ)AAI30637513
035 $a AAI30637513
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Zhang, Xinwei. $3 1468454
245 1 0 $a Advanced Machine Learning Methods : $b Semi-Supervised Learning, Multi-Category Classification, and Energy-Based Models.
264 0 $c 2023
300 $a 1 online resource (207 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: 85-05, Section: B.
500 $a Advisor: Tan, Zhiqiang.
502 $a Thesis (Ph.D.)--Rutgers The State University of New Jersey, School of Graduate Studies, 2023.
504 $a Includes bibliographical references
520 $a This dissertation explores diverse aspects of machine learning with a particular focus on semi-supervised learning, multi-category classification, and energy-based models. The work begins by extending the statistical equivalence between logistic regression and exponential tilt modeling, leading to the development of a novel semi-supervised logistic learning method. Enhanced prediction accuracy is achieved by leveraging both labeled and unlabeled data. Next, we delve into the construction of loss functions and the establishment of corresponding regret bounds for multi-category classification. Here, new general representations of losses are derived, leading to the discovery of new hinge-like convex losses and multi-class proper scoring rules. In the final part, we introduce diffusion data and propose to learn a joint energy-based model, through persistent training, which simultaneously achieves long-run stability, post-training image generation, and superior out-of-distribution detection. Overall, this dissertation provides valuable insights and advancements in the field of machine learning, revealing connections between different methods and contributing to the improvement of existing methods and theories.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Statistics. $3 556824
650 4 $a Computer science. $3 573171
650 4 $a Information technology. $3 559429
653 $a Energy-based models
653 $a Multicategory classification
653 $a Semi-supervised learning
653 $a Machine learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0463
690 $a 0984
690 $a 0489
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Rutgers The State University of New Jersey, School of Graduate Studies. $b Statistics and Biostatistics. $3 1241305
773 0 $t Dissertations Abstracts International $g 85-05B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30637513 $z click for full text (PQDT)