東華大學圖書館 |

Neural Network Architecture Optimization Using Reinforcement Learning.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Neural Network Architecture Optimization Using Reinforcement Learning./
作者:	Vadhera, Raghav.
面頁冊數:	1 online resource (215 pages)
附註:	Source: Dissertations Abstracts International, Volume: 84-12, Section: B.
Contained By:	Dissertations Abstracts International84-12B.
標題:	Mass communications. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30593108click for full text (PQDT)
ISBN:	9798379715144

Neural Network Architecture Optimization Using Reinforcement Learning.
Vadhera, Raghav.

Neural Network Architecture Optimization Using Reinforcement Learning. - 1 online resource (215 pages)

Source: Dissertations Abstracts International, Volume: 84-12, Section: B.

Thesis (Ph.D.)--The University of Texas at Arlington, 2023.

Includes bibliographical references

Deep learning has emerged as an increasingly valuable tool, employed across a myriad of applications. However, the intricacies of deep learning systems, stemming from their sensitivity to specific network architectures, have rendered them challenging for non-experts to harness, thus highlighting the need for automatic network architecture optimization. Prior research predominantly optimizes a network for a single problem through architecture search, necessitating extensive training of various architectures during optimization.To tackle this issue and unlock the potential for transferability across tasks, this dissertation presents a novel approach that employs Reinforcement Learning to develop a network optimization policy based on an abstract problem and architecture embedding. This approach enables the optimization of networks for novel problems without the burden of excessive additional training. Leveraging policy learning and an abstract problem embedding, the method facilitates the transfer of the policy across problems by capturing essential characteristics of the network domain and target task that permit the approach to optimize the networks for new challenges based on characteristics learned from previous problems.Initial evaluations of this method's capabilities were conducted using a standard classification problem, demonstrating its effectiveness in optimizing architectures for a specific target problem within a given range of fully connected networks. Subsequent experiments were performed using a variety of complex problems, further showcasing the approach's capabilities. To address these more complex networks, Siamese networks were employed to establish a coherent embedding of the network architecture space. In conjunction with a problem-specific feature vector, which captures the intricacies of the problem, the Reinforcement Learning agent was able to acquire a transferable policy for deriving high-performing network architectures across a spectrum of problems.Experiments performed in this dissertation specifically reveal that the proposed system successfully learns an embedding space and policy that can derive and optimize network architectures nearing optimality, even for unencountered problems. Multiple datasets, each possessing unique feature vectors representing distinct characteristics entities or problems, were utilized to facilitate the optimization of one problem at a time. A random initial policy was employed to construct trajectories in the embedding space during training. To assess the performance and functionality of various network components, a series of pre-training steps were undertaken, focusing on distinct components and examining the outcomes prior to training subsequent components.Building upon these foundations, the dissertation takes initial steps to examine the scalability of the method to larger and more intricate network architectures with the intent of broadening its applicability across a diverse array of problem domains.To validate the generalizability of the learned policies, the dissertation examines their performance on real-world problems, spanning various industries and domains, including healthcare, finance, sports, human psychology and auto. These case studies aim to demonstrate the practical utility of the proposed approach in addressing real-world challenges and uncover potential areas for further refinement and improvement.In addition to these empirical investigations, the dissertation discusses the theoretical underpinnings of the method, examining the convergence properties, stability, and robustness of the learned policies. These investigations provide valuable insights into the factors that influence policy transferability and optimization performance across diverse problem domains, offering guidance for future research in the field of deep learning and network architecture optimization.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798379715144Subjects--Topical Terms:

3422380
Mass communications.
Subjects--Index Terms:

Reinforcement learningIndex Terms--Genre/Form:

542853
Electronic books.

Neural Network Architecture Optimization Using Reinforcement Learning.
LDR:05356nmm a2200433K 4500 001 2364845
005 20231212064435.5
006 m o d
007 cr mn ---uuuuu
008 241011s2023 xx obm 000 0 eng d
020 $a 9798379715144
035 $a (MiAaPQ)AAI30593108
035 $a (MiAaPQ)2502vireo3037Vadhera
035 $a AAI30593108
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Vadhera, Raghav. $3 3705668
245 1 0 $a Neural Network Architecture Optimization Using Reinforcement Learning.
264 0 $c 2023
300 $a 1 online resource (215 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: 84-12, Section: B.
500 $a Advisor: Huber, Manfred.
502 $a Thesis (Ph.D.)--The University of Texas at Arlington, 2023.
504 $a Includes bibliographical references
520 $a Deep learning has emerged as an increasingly valuable tool, employed across a myriad of applications. However, the intricacies of deep learning systems, stemming from their sensitivity to specific network architectures, have rendered them challenging for non-experts to harness, thus highlighting the need for automatic network architecture optimization. Prior research predominantly optimizes a network for a single problem through architecture search, necessitating extensive training of various architectures during optimization.To tackle this issue and unlock the potential for transferability across tasks, this dissertation presents a novel approach that employs Reinforcement Learning to develop a network optimization policy based on an abstract problem and architecture embedding. This approach enables the optimization of networks for novel problems without the burden of excessive additional training. Leveraging policy learning and an abstract problem embedding, the method facilitates the transfer of the policy across problems by capturing essential characteristics of the network domain and target task that permit the approach to optimize the networks for new challenges based on characteristics learned from previous problems.Initial evaluations of this method's capabilities were conducted using a standard classification problem, demonstrating its effectiveness in optimizing architectures for a specific target problem within a given range of fully connected networks. Subsequent experiments were performed using a variety of complex problems, further showcasing the approach's capabilities. To address these more complex networks, Siamese networks were employed to establish a coherent embedding of the network architecture space. In conjunction with a problem-specific feature vector, which captures the intricacies of the problem, the Reinforcement Learning agent was able to acquire a transferable policy for deriving high-performing network architectures across a spectrum of problems.Experiments performed in this dissertation specifically reveal that the proposed system successfully learns an embedding space and policy that can derive and optimize network architectures nearing optimality, even for unencountered problems. Multiple datasets, each possessing unique feature vectors representing distinct characteristics entities or problems, were utilized to facilitate the optimization of one problem at a time. A random initial policy was employed to construct trajectories in the embedding space during training. To assess the performance and functionality of various network components, a series of pre-training steps were undertaken, focusing on distinct components and examining the outcomes prior to training subsequent components.Building upon these foundations, the dissertation takes initial steps to examine the scalability of the method to larger and more intricate network architectures with the intent of broadening its applicability across a diverse array of problem domains.To validate the generalizability of the learned policies, the dissertation examines their performance on real-world problems, spanning various industries and domains, including healthcare, finance, sports, human psychology and auto. These case studies aim to demonstrate the practical utility of the proposed approach in addressing real-world challenges and uncover potential areas for further refinement and improvement.In addition to these empirical investigations, the dissertation discusses the theoretical underpinnings of the method, examining the convergence properties, stability, and robustness of the learned policies. These investigations provide valuable insights into the factors that influence policy transferability and optimization performance across diverse problem domains, offering guidance for future research in the field of deep learning and network architecture optimization.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Mass communications. $3 3422380
650 4 $a Computer science. $3 523869
653 $a Reinforcement learning
653 $a Deep learning
653 $a NAO
653 $a TD3
653 $a Agent
653 $a Critic
653 $a Siamese
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0800
690 $a 0984
690 $a 0708
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a The University of Texas at Arlington. $b Computer Science and Engineering. $3 2104073
773 0 $t Dissertations Abstracts International $g 84-12B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30593108 $z click for full text (PQDT)