國立虎尾科技大學 |

Understanding and Improving Reliability of Inference Dynamics in Deep Neural Networks.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Understanding and Improving Reliability of Inference Dynamics in Deep Neural Networks./
作者:	Yujia Huang.
面頁冊數:	1 online resource (165 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
Contained By:	Dissertations Abstracts International85-11B.
標題:	Computer engineering. -
電子資源:	click for full text (PQDT)
ISBN:	9798382237947

Understanding and Improving Reliability of Inference Dynamics in Deep Neural Networks.
Yujia Huang.

Understanding and Improving Reliability of Inference Dynamics in Deep Neural Networks. - 1 online resource (165 pages)

Source: Dissertations Abstracts International, Volume: 85-11, Section: B.

Thesis (Ph.D.)--California Institute of Technology, 2024.

Includes bibliographical references

Reliability is a crucial aspect for the successful deployment of deep learning systems across various domains. In generative modeling, it is essential to create content that adheres to specific rules. In the field of control, ensuring that robots operate safely without falling or entering hazardous areas is paramount. Similarly, in visual perception, the robustness of perception results against perturbations are vital.In this thesis, we explore the reliability of inference dynamics in deep neural networks such as ResNet, neural Ordinary Differential Equations (ODEs), and diffusion models. We begin by examining the inference dynamics in standard networks with a discrete sequence of hidden layers, applying self-consistency and local Lipschitz bounds to enhance robustness against input perturbations. Our exploration then extends to neural ODEs, where the neural network specifies a vector field that continuously transforms the state. We employ forward invariance to achieve robustness, marking the first instance of training neural ODE policies with non-vacuous certified guarantees. The focus shifts next to diffusion models and their inference processes, particularly in adhering to symbolic constraints. For this, we introduce a novel sampling algorithm inspired by stochastic control principles. This algorithm not only guides these models in generating rule-specific content but also sets a new benchmark in symbolic music generation. Our work offers a cohesive understanding of inference dynamics in various deep learning architectures and propose new algorithms to significantly improve their reliability.

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

Mode of access: World Wide Web

ISBN: 9798382237947Subjects--Topical Terms:

569006
Computer engineering.
Subjects--Index Terms:

Inference dyanmicsIndex Terms--Genre/Form:

554714
Electronic books.

Understanding and Improving Reliability of Inference Dynamics in Deep Neural Networks.
LDR:02904ntm a22003497 4500 001 1148396
005 20240924101915.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798382237947
035 $a (MiAaPQ)AAI31266773
035 $a (MiAaPQ)16314
035 $a AAI31266773
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Yujia Huang. $3 1474349
245 1 0 $a Understanding and Improving Reliability of Inference Dynamics in Deep Neural Networks.
264 0 $c 2024
300 $a 1 online resource (165 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: 85-11, Section: B.
500 $a Advisor: Yisong Yue.
502 $a Thesis (Ph.D.)--California Institute of Technology, 2024.
504 $a Includes bibliographical references
520 $a Reliability is a crucial aspect for the successful deployment of deep learning systems across various domains. In generative modeling, it is essential to create content that adheres to specific rules. In the field of control, ensuring that robots operate safely without falling or entering hazardous areas is paramount. Similarly, in visual perception, the robustness of perception results against perturbations are vital.In this thesis, we explore the reliability of inference dynamics in deep neural networks such as ResNet, neural Ordinary Differential Equations (ODEs), and diffusion models. We begin by examining the inference dynamics in standard networks with a discrete sequence of hidden layers, applying self-consistency and local Lipschitz bounds to enhance robustness against input perturbations. Our exploration then extends to neural ODEs, where the neural network specifies a vector field that continuously transforms the state. We employ forward invariance to achieve robustness, marking the first instance of training neural ODE policies with non-vacuous certified guarantees. The focus shifts next to diffusion models and their inference processes, particularly in adhering to symbolic constraints. For this, we introduce a novel sampling algorithm inspired by stochastic control principles. This algorithm not only guides these models in generating rule-specific content but also sets a new benchmark in symbolic music generation. Our work offers a cohesive understanding of inference dynamics in various deep learning architectures and propose new algorithms to significantly improve their reliability.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer engineering. $3 569006
653 $a Inference dyanmics
653 $a Neural networks
655 7 $a Electronic books. $2 local $3 554714
690 $a 0464
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a California Institute of Technology. $b Engineering and Applied Science. $3 1437764
773 0 $t Dissertations Abstracts International $g 85-11B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31266773 $z click for full text (PQDT)