國立高雄大學圖資館 |

Inductive biases in machine learning for robotics and control

Record Type:	Electronic resources : Monograph/item
Title/Author:	Inductive biases in machine learning for robotics and controlby Michael Lutter.
Author:	Lutter, Michael.
Published:	Cham :Springer Nature Switzerland :2023.
Description:	xv, 119 p. :ill. (some col.), digital ;24 cm.
Contained By:	Springer Nature eBook
Subject:	Machine learning.
Online resource:	https://doi.org/10.1007/978-3-031-37832-4
ISBN:	9783031378324$q(electronic bk.)

Inductive biases in machine learning for robotics and control
Lutter, Michael.

Inductive biases in machine learning for robotics and control[electronic resource] /by Michael Lutter. - Cham :Springer Nature Switzerland :2023. - xv, 119 p. :ill. (some col.), digital ;24 cm. - Springer tracts in advanced robotics,v. 1561610-742X ;. - Springer tracts in advanced robotics ;v. 13.

Introduction -- A Differentiable Newton-Euler Algorithm for Real-World Robotics -- Combining Physics and Deep Learning for Continuous-Time Dynamics Models -- Continuous-Time Fitted Value Iteration for Robust Policies -- Conclusion.

One important robotics problem is "How can one program a robot to perform a task"? Classical robotics solves this problem by manually engineering modules for state estimation, planning, and control. In contrast, robot learning solely relies on black-box models and data. This book shows that these two approaches of classical engineering and black-box machine learning are not mutually exclusive. To solve tasks with robots, one can transfer insights from classical robotics to deep networks and obtain better learning algorithms for robotics and control. To highlight that incorporating existing knowledge as inductive biases in machine learning algorithms improves performance, this book covers different approaches for learning dynamics models and learning robust control policies. The presented algorithms leverage the knowledge of Newtonian Mechanics, Lagrangian Mechanics as well as the Hamilton-Jacobi-Isaacs differential equation as inductive bias and are evaluated on physical robots.

ISBN: 9783031378324$q(electronic bk.)

Standard No.: 10.1007/978-3-031-37832-4doiSubjects--Topical Terms:

188639
Machine learning.

LC Class. No.: Q325.5

Dewey Class. No.: 006.31

Inductive biases in machine learning for robotics and control
LDR:02300nmm a2200337 a 4500 001 648721
003 DE-He213
005 20230801002656.0
006 m d
007 cr nn 008maaau
008 240116s2023 sz s 0 eng d
020 $a 9783031378324$q(electronic bk.)
020 $a 9783031378317$q(paper)
024 7 $a 10.1007/978-3-031-37832-4 $2 doi
035 $a 978-3-031-37832-4
040 $a GP $c GP
041 0 $a eng
050 4 $a Q325.5
072 7 $a TJFM $2 bicssc
072 7 $a TEC004000 $2 bisacsh
072 7 $a TJFM $2 thema
082 0 4 $a 006.31 $2 23
090 $a Q325.5 $b .L974 2023
100 1 $a Lutter, Michael. $3 958520
245 1 0 $a Inductive biases in machine learning for robotics and control $h [electronic resource] / $c by Michael Lutter.
260 $a Cham : $b Springer Nature Switzerland : $b Imprint: Springer, $c 2023.
300 $a xv, 119 p. : $b ill. (some col.), digital ; $c 24 cm.
490 1 $a Springer tracts in advanced robotics, $x 1610-742X ; $v v. 156
505 0 $a Introduction -- A Differentiable Newton-Euler Algorithm for Real-World Robotics -- Combining Physics and Deep Learning for Continuous-Time Dynamics Models -- Continuous-Time Fitted Value Iteration for Robust Policies -- Conclusion.
520 $a One important robotics problem is "How can one program a robot to perform a task"? Classical robotics solves this problem by manually engineering modules for state estimation, planning, and control. In contrast, robot learning solely relies on black-box models and data. This book shows that these two approaches of classical engineering and black-box machine learning are not mutually exclusive. To solve tasks with robots, one can transfer insights from classical robotics to deep networks and obtain better learning algorithms for robotics and control. To highlight that incorporating existing knowledge as inductive biases in machine learning algorithms improves performance, this book covers different approaches for learning dynamics models and learning robust control policies. The presented algorithms leverage the knowledge of Newtonian Mechanics, Lagrangian Mechanics as well as the Hamilton-Jacobi-Isaacs differential equation as inductive bias and are evaluated on physical robots.
650 0 $a Machine learning. $3 188639
650 0 $a Robotics. $3 181952
650 0 $a Discrimination. $3 201243
650 1 4 $a Control, Robotics, Automation. $3 914981
650 2 4 $a Computational Intelligence. $3 338479
650 2 4 $a Control and Systems Theory. $3 825946
710 2 $a SpringerLink (Online service) $3 273601
773 0 $t Springer Nature eBook
830 0 $a Springer tracts in advanced robotics ; $v v. 13 $3 452791
856 4 0 $u https://doi.org/10.1007/978-3-031-37832-4
950 $a Intelligent Technologies and Robotics (SpringerNature-42732)