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

Imagination-Augmented Deep Reinforcement Learning for Robotic Applications.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Imagination-Augmented Deep Reinforcement Learning for Robotic Applications./
作者:	Thabet, Mohammad.
面頁冊數:	1 online resource (118 pages)
附註:	Source: Dissertations Abstracts International, Volume: 84-02, Section: B.
Contained By:	Dissertations Abstracts International84-02B.
標題:	Kinematics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29216640click for full text (PQDT)
ISBN:	9798841522195

Imagination-Augmented Deep Reinforcement Learning for Robotic Applications.
Thabet, Mohammad.

Imagination-Augmented Deep Reinforcement Learning for Robotic Applications. - 1 online resource (118 pages)

Source: Dissertations Abstracts International, Volume: 84-02, Section: B.

Thesis (Ph.D.)--The University of Manchester (United Kingdom), 2022.

Includes bibliographical references

Deep reinforcement learning (RL) has recently emerged as a powerful technique that allows agents to solve complex sequential tasks. Its application in the field of robotics however has been held back by the impracticality of the enormous amount of interaction data it requires. Collecting data with a physical robot is usually prohibitively costly, prompting the need for more sample-efficient reinforcement learning algorithms. This thesis aims to develop an architecture that drastically improves the sample efficiency of RL and leads to faster learning with less data. The architecture incorporates a mechanism that mimics imagination in humans into model-free learning, allowing agents to simulate scenarios internally to lessen the need for actual interaction with the environment. In this model-assisted setting, an agent learns a stochastic environment model on-line from experience simultaneously with a policy. A variational autoencoder (VAE) is used to compress visual input into abstract representations in the latent space, and a mixture density network (MDN) is used to learn a forward model in this latent space. The agent then uses the learned model to generate imaginary rollouts to augment real data, which is then used to train a controller in an RL context. Uncertainty in the model predictions is estimated using Monte-Carlo dropout to limit the use of imaginary data, preventing the agent from using erroneous model predictions in learning. The thesis presents experiments involving human-robot interaction scenarios in an RL setting to verify the viability of the approach. The first experiment serves as a proof of concept and involved an agent learning a pick-and-place task based on gestures by a human. The second experiment was designed to demonstrate the advantages of the approach and involved a robot learning to solve a puzzle based on gestures. Results show that the proposed imagination-augmented agents perform significantly better than baseline agents when data is scarce, proving the efficacy of the approach in increasing sample efficiency.

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

Mode of access: World Wide Web

ISBN: 9798841522195Subjects--Topical Terms:

571109
Kinematics.
Index Terms--Genre/Form:

542853
Electronic books.

Imagination-Augmented Deep Reinforcement Learning for Robotic Applications.
LDR:03488nmm a2200397K 4500 001 2354272
005 20230403071205.5
006 m o d
007 cr mn ---uuuuu
008 241011s2022 xx obm 000 0 eng d
020 $a 9798841522195
035 $a (MiAaPQ)AAI29216640
035 $a (MiAaPQ)Manchester_UKac51471b-b3a5-4a8c-8fd4-24a492a19977
035 $a AAI29216640
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Thabet, Mohammad. $3 3694621
245 1 0 $a Imagination-Augmented Deep Reinforcement Learning for Robotic Applications.
264 0 $c 2022
300 $a 1 online resource (118 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: 84-02, Section: B.
500 $a Advisor: Shapiro, Jonathan; Cangelosi, Angelo.
502 $a Thesis (Ph.D.)--The University of Manchester (United Kingdom), 2022.
504 $a Includes bibliographical references
520 $a Deep reinforcement learning (RL) has recently emerged as a powerful technique that allows agents to solve complex sequential tasks. Its application in the field of robotics however has been held back by the impracticality of the enormous amount of interaction data it requires. Collecting data with a physical robot is usually prohibitively costly, prompting the need for more sample-efficient reinforcement learning algorithms. This thesis aims to develop an architecture that drastically improves the sample efficiency of RL and leads to faster learning with less data. The architecture incorporates a mechanism that mimics imagination in humans into model-free learning, allowing agents to simulate scenarios internally to lessen the need for actual interaction with the environment. In this model-assisted setting, an agent learns a stochastic environment model on-line from experience simultaneously with a policy. A variational autoencoder (VAE) is used to compress visual input into abstract representations in the latent space, and a mixture density network (MDN) is used to learn a forward model in this latent space. The agent then uses the learned model to generate imaginary rollouts to augment real data, which is then used to train a controller in an RL context. Uncertainty in the model predictions is estimated using Monte-Carlo dropout to limit the use of imaginary data, preventing the agent from using erroneous model predictions in learning. The thesis presents experiments involving human-robot interaction scenarios in an RL setting to verify the viability of the approach. The first experiment serves as a proof of concept and involved an agent learning a pick-and-place task based on gestures by a human. The second experiment was designed to demonstrate the advantages of the approach and involved a robot learning to solve a puzzle based on gestures. Results show that the proposed imagination-augmented agents perform significantly better than baseline agents when data is scarce, proving the efficacy of the approach in increasing sample efficiency.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Kinematics. $3 571109
650 4 $a Neurons. $3 588699
650 4 $a Deep learning. $3 3554982
650 4 $a Back propagation. $3 3681810
650 4 $a Success. $3 518195
650 4 $a Power. $3 518736
650 4 $a Decision making. $3 517204
650 4 $a Neural networks. $3 677449
650 4 $a Robots. $3 529507
650 4 $a Imagination. $3 542038
650 4 $a Algorithms. $3 536374
650 4 $a Expected values. $3 3563993
650 4 $a Cognition & reasoning. $3 3556293
650 4 $a Robotics. $3 519753
650 4 $a Cognitive psychology. $3 523881
650 4 $a Computer science. $3 523869
650 4 $a Mechanics. $3 525881
650 4 $a Physics. $3 516296
650 4 $a Psychology. $3 519075
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0771
690 $a 0800
690 $a 0633
690 $a 0984
690 $a 0346
690 $a 0605
690 $a 0621
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a The University of Manchester (United Kingdom). $3 3422292
773 0 $t Dissertations Abstracts International $g 84-02B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29216640 $z click for full text (PQDT)