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

A Picture of the Energy Landscape of...

Chaudhari, Pratik Anil.

FindBook

Google Book

Amazon

博客來

A Picture of the Energy Landscape of Deep Neural Networks.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	A Picture of the Energy Landscape of Deep Neural Networks./
作者:	Chaudhari, Pratik Anil.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2018,
面頁冊數:	175 p.
附註:	Source: Dissertations Abstracts International, Volume: 80-03, Section: B.
Contained By:	Dissertations Abstracts International80-03B.
標題:	Statistical physics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10843897
ISBN:	9780438291706

A Picture of the Energy Landscape of Deep Neural Networks.
Chaudhari, Pratik Anil.

A Picture of the Energy Landscape of Deep Neural Networks. - Ann Arbor : ProQuest Dissertations & Theses, 2018 - 175 p.

Source: Dissertations Abstracts International, Volume: 80-03, Section: B.

Thesis (Ph.D.)--University of California, Los Angeles, 2018.

This item must not be added to any third party search indexes.

This thesis characterizes the training process of deep neural networks. We are driven by two apparent paradoxes. First, optimizing a non-convex function such as the loss function of a deep network should be extremely hard, yet rudimentary algorithms like stochastic gradient descent are phenomenally successful at this. Second, over-parametrized models are expected to perform poorly on new data, yet large deep networks with millions of parameters achieve spectacular generalization performance. We build upon tools from two main areas to make progress on these questions: statistical physics and a continuous-time point-of-view of optimization. The former has been popular in the study of machine learning in the past and has been rejuvenated in recent years due to the strong correlation of empirical properties of modern deep networks with existing, older analytical results. The latter, i.e., modeling stochastic first-order algorithms as continuous-time stochastic processes, gives access to powerful tools from the theory of partial differential equations, optimal transportation and non-equilibrium thermodynamics. The confluence of these ideas leads to fundamental theoretical insights that explain observed phenomena in deep learning as well as the development of state-of-the-art algorithms for training deep networks.

ISBN: 9780438291706Subjects--Topical Terms:

536281
Statistical physics.

A Picture of the Energy Landscape of Deep Neural Networks.
LDR:02497nmm a2200349 4500 001 2210917
005 20191126113855.5
008 201008s2018 ||||||||||||||||| ||eng d
020 $a 9780438291706
035 $a (MiAaPQ)AAI10843897
035 $a (MiAaPQ)ucla:17157
035 $a AAI10843897
040 $a MiAaPQ $c MiAaPQ
100 1 $a Chaudhari, Pratik Anil. $3 3438061
245 1 0 $a A Picture of the Energy Landscape of Deep Neural Networks.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2018
300 $a 175 p.
500 $a Source: Dissertations Abstracts International, Volume: 80-03, Section: B.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Soatto, Stefano.
502 $a Thesis (Ph.D.)--University of California, Los Angeles, 2018.
506 $a This item must not be added to any third party search indexes.
506 $a This item must not be sold to any third party vendors.
520 $a This thesis characterizes the training process of deep neural networks. We are driven by two apparent paradoxes. First, optimizing a non-convex function such as the loss function of a deep network should be extremely hard, yet rudimentary algorithms like stochastic gradient descent are phenomenally successful at this. Second, over-parametrized models are expected to perform poorly on new data, yet large deep networks with millions of parameters achieve spectacular generalization performance. We build upon tools from two main areas to make progress on these questions: statistical physics and a continuous-time point-of-view of optimization. The former has been popular in the study of machine learning in the past and has been rejuvenated in recent years due to the strong correlation of empirical properties of modern deep networks with existing, older analytical results. The latter, i.e., modeling stochastic first-order algorithms as continuous-time stochastic processes, gives access to powerful tools from the theory of partial differential equations, optimal transportation and non-equilibrium thermodynamics. The confluence of these ideas leads to fundamental theoretical insights that explain observed phenomena in deep learning as well as the development of state-of-the-art algorithms for training deep networks.
590 $a School code: 0031.
650 4 $a Statistical physics. $3 536281
650 4 $a Applied Mathematics. $3 1669109
650 4 $a Artificial intelligence. $3 516317
690 $a 0217
690 $a 0364
690 $a 0800
710 2 $a University of California, Los Angeles. $b Computer Science 0201. $3 2049859
773 0 $t Dissertations Abstracts International $g 80-03B.
790 $a 0031
791 $a Ph.D.
792 $a 2018
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10843897