東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

FindBook

Google Book

Amazon

博客來

Accelerating Machine Learning via Multi-Objective Optimization.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Accelerating Machine Learning via Multi-Objective Optimization./
作者:	Lim, Robert.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2021,
面頁冊數:	147 p.
附註:	Source: Dissertations Abstracts International, Volume: 83-05, Section: B.
Contained By:	Dissertations Abstracts International83-05B.
標題:	Computer science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28722564
ISBN:	9798492727826

Accelerating Machine Learning via Multi-Objective Optimization.
Lim, Robert.

Accelerating Machine Learning via Multi-Objective Optimization. - Ann Arbor : ProQuest Dissertations & Theses, 2021 - 147 p.

Source: Dissertations Abstracts International, Volume: 83-05, Section: B.

Thesis (Ph.D.)--University of Oregon, 2021.

This item must not be sold to any third party vendors.

This dissertation work presents various approaches toward accelerating training of deep neural networks with the use of high-performance computing resources, while balancing learning and systems utilization objectives. Acceleration of machine learning is formulated as a multi-objective optimization problem that seeks to satisfy multiple objectives, based on its respective constraints. In machine learning, the objective is to strive for a model that has high accuracy, while eliminating false positives and generalizing beyond the training set. For systems execution performance, maximizing utilization of the underlying hardware resources within compute and power budgets are constraints that bound the problem. In both scenarios, the search space is combinatorial and contains multiple local minima that in many cases satisfies the global optimum. This dissertation work addresses the search of solutions in both performance tuning and neural network training. Specifically, subgraph matching is proposed to bound the search problem and provide heuristics that guide the solver toward the optimal solution. Mixed precision operations is also proposed for solving systems of linear equations and for training neural networks for image classification for evaluating the stability and robustness of the operations. Use cases are presented with CUDA performance tuning and neural network training, demonstrating the effectiveness of the proposed technique. The experiments were carried out on single and multi-node GPU clusters, and reveals opportunities for further exploration in this critical hardware/software co-design space of accelerated machine learning.

ISBN: 9798492727826Subjects--Topical Terms:

523869
Computer science.
Subjects--Index Terms:

Accelerated machine learning

Accelerating Machine Learning via Multi-Objective Optimization.
LDR:02774nmm a2200349 4500 001 2352151
005 20221118093824.5
008 241004s2021 ||||||||||||||||| ||eng d
020 $a 9798492727826
035 $a (MiAaPQ)AAI28722564
035 $a AAI28722564
040 $a MiAaPQ $c MiAaPQ
100 1 $a Lim, Robert. $3 3412564
245 1 0 $a Accelerating Machine Learning via Multi-Objective Optimization.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2021
300 $a 147 p.
500 $a Source: Dissertations Abstracts International, Volume: 83-05, Section: B.
500 $a Advisor: Malony, Allen.
502 $a Thesis (Ph.D.)--University of Oregon, 2021.
506 $a This item must not be sold to any third party vendors.
520 $a This dissertation work presents various approaches toward accelerating training of deep neural networks with the use of high-performance computing resources, while balancing learning and systems utilization objectives. Acceleration of machine learning is formulated as a multi-objective optimization problem that seeks to satisfy multiple objectives, based on its respective constraints. In machine learning, the objective is to strive for a model that has high accuracy, while eliminating false positives and generalizing beyond the training set. For systems execution performance, maximizing utilization of the underlying hardware resources within compute and power budgets are constraints that bound the problem. In both scenarios, the search space is combinatorial and contains multiple local minima that in many cases satisfies the global optimum. This dissertation work addresses the search of solutions in both performance tuning and neural network training. Specifically, subgraph matching is proposed to bound the search problem and provide heuristics that guide the solver toward the optimal solution. Mixed precision operations is also proposed for solving systems of linear equations and for training neural networks for image classification for evaluating the stability and robustness of the operations. Use cases are presented with CUDA performance tuning and neural network training, demonstrating the effectiveness of the proposed technique. The experiments were carried out on single and multi-node GPU clusters, and reveals opportunities for further exploration in this critical hardware/software co-design space of accelerated machine learning.
590 $a School code: 0171.
650 4 $a Computer science. $3 523869
650 4 $a Artificial intelligence. $3 516317
650 4 $a Information science. $3 554358
653 $a Accelerated machine learning
653 $a Multi-objective optimization
653 $a Neural network
690 $a 0984
690 $a 0723
690 $a 0800
710 2 $a University of Oregon. $b Department of Computer and Information Science. $3 3347138
773 0 $t Dissertations Abstracts International $g 83-05B.
790 $a 0171
791 $a Ph.D.
792 $a 2021
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28722564