Measuring progress in robotics: benchmarking and the ‘measure-target confusion’

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

Abstract

While it is often said that in order to qualify as a true science robotics should aspire to reproducible and measurable results that allow benchmarking, I argue that a focus on benchmarking will be a hindrance for progress. Several academic disciplines that have been led into pursuing only reproducible and measurable ‘scientific’ results—robotics should be careful not to fall into that trap. Results that can be benchmarked must be specific and context-dependent, but robotics targets whole complex systems independently of a specific context—so working towards progress on the technical measure risks missing that target. It would constitute aiming for the measure rather than the target: what I call ‘measure-target confusion’. The role of benchmarking in robotics shows that the more general problem to measure progress towards more intelligent machines will not be solved by technical benchmarks; we need a balanced approach with technical benchmarks, real-life testing and qualitative judgment.

LanguageEnglish
Title of host publicationMetrics of Sensory Motor Coordination and Integration in Robots and Animals
Subtitle of host publicationHow to Measure the Success of Bioinspired Solutions with Respect to their Natural Models, and Against More ‘Artificial’ Solutions?
EditorsFabio Bonsignorio, Elena Messina, Angel P. del Pobil, John Hallam
Place of PublicationCham
PublisherSpringer
Pages169-179
Number of pages11
ISBN (Electronic)978-3-030-14126-4
ISBN (Print)978-3-030-14124-0
DOIs
StatePublished - 1 Jan 2020

Publication series

NameCognitive Systems Monographs
Volume36
ISSN (Print)1867-4925
ISSN (Electronic)1867-4933

Fingerprint

Benchmarking
Robotics
Large scale systems
Testing

Cite this

Müller, V. C. (2020). Measuring progress in robotics: benchmarking and the ‘measure-target confusion’. In F. Bonsignorio, E. Messina, A. P. del Pobil, & J. Hallam (Eds.), Metrics of Sensory Motor Coordination and Integration in Robots and Animals: How to Measure the Success of Bioinspired Solutions with Respect to their Natural Models, and Against More ‘Artificial’ Solutions? (pp. 169-179). (Cognitive Systems Monographs; Vol. 36). Cham: Springer. DOI: 10.1007/978-3-030-14126-4_9
Müller, Vincent C./ Measuring progress in robotics : benchmarking and the ‘measure-target confusion’. Metrics of Sensory Motor Coordination and Integration in Robots and Animals: How to Measure the Success of Bioinspired Solutions with Respect to their Natural Models, and Against More ‘Artificial’ Solutions?. editor / Fabio Bonsignorio ; Elena Messina ; Angel P. del Pobil ; John Hallam. Cham : Springer, 2020. pp. 169-179 (Cognitive Systems Monographs).
@inbook{455802cbbbd048e68b729d20440e2d8c,
title = "Measuring progress in robotics: benchmarking and the ‘measure-target confusion’",
abstract = "While it is often said that in order to qualify as a true science robotics should aspire to reproducible and measurable results that allow benchmarking, I argue that a focus on benchmarking will be a hindrance for progress. Several academic disciplines that have been led into pursuing only reproducible and measurable ‘scientific’ results—robotics should be careful not to fall into that trap. Results that can be benchmarked must be specific and context-dependent, but robotics targets whole complex systems independently of a specific context—so working towards progress on the technical measure risks missing that target. It would constitute aiming for the measure rather than the target: what I call ‘measure-target confusion’. The role of benchmarking in robotics shows that the more general problem to measure progress towards more intelligent machines will not be solved by technical benchmarks; we need a balanced approach with technical benchmarks, real-life testing and qualitative judgment.",
author = "M{\"u}ller, {Vincent C.}",
year = "2020",
month = "1",
day = "1",
doi = "10.1007/978-3-030-14126-4_9",
language = "English",
isbn = "978-3-030-14124-0",
series = "Cognitive Systems Monographs",
publisher = "Springer",
pages = "169--179",
editor = "Fabio Bonsignorio and Elena Messina and {del Pobil}, {Angel P.} and John Hallam",
booktitle = "Metrics of Sensory Motor Coordination and Integration in Robots and Animals",
address = "Germany",

}

Müller, VC 2020, Measuring progress in robotics: benchmarking and the ‘measure-target confusion’. in F Bonsignorio, E Messina, AP del Pobil & J Hallam (eds), Metrics of Sensory Motor Coordination and Integration in Robots and Animals: How to Measure the Success of Bioinspired Solutions with Respect to their Natural Models, and Against More ‘Artificial’ Solutions?. Cognitive Systems Monographs, vol. 36, Springer, Cham, pp. 169-179. DOI: 10.1007/978-3-030-14126-4_9

Measuring progress in robotics : benchmarking and the ‘measure-target confusion’. / Müller, Vincent C.

Metrics of Sensory Motor Coordination and Integration in Robots and Animals: How to Measure the Success of Bioinspired Solutions with Respect to their Natural Models, and Against More ‘Artificial’ Solutions?. ed. / Fabio Bonsignorio; Elena Messina; Angel P. del Pobil; John Hallam. Cham : Springer, 2020. p. 169-179 (Cognitive Systems Monographs; Vol. 36).

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

TY - CHAP

T1 - Measuring progress in robotics

T2 - benchmarking and the ‘measure-target confusion’

AU - Müller,Vincent C.

PY - 2020/1/1

Y1 - 2020/1/1

N2 - While it is often said that in order to qualify as a true science robotics should aspire to reproducible and measurable results that allow benchmarking, I argue that a focus on benchmarking will be a hindrance for progress. Several academic disciplines that have been led into pursuing only reproducible and measurable ‘scientific’ results—robotics should be careful not to fall into that trap. Results that can be benchmarked must be specific and context-dependent, but robotics targets whole complex systems independently of a specific context—so working towards progress on the technical measure risks missing that target. It would constitute aiming for the measure rather than the target: what I call ‘measure-target confusion’. The role of benchmarking in robotics shows that the more general problem to measure progress towards more intelligent machines will not be solved by technical benchmarks; we need a balanced approach with technical benchmarks, real-life testing and qualitative judgment.

AB - While it is often said that in order to qualify as a true science robotics should aspire to reproducible and measurable results that allow benchmarking, I argue that a focus on benchmarking will be a hindrance for progress. Several academic disciplines that have been led into pursuing only reproducible and measurable ‘scientific’ results—robotics should be careful not to fall into that trap. Results that can be benchmarked must be specific and context-dependent, but robotics targets whole complex systems independently of a specific context—so working towards progress on the technical measure risks missing that target. It would constitute aiming for the measure rather than the target: what I call ‘measure-target confusion’. The role of benchmarking in robotics shows that the more general problem to measure progress towards more intelligent machines will not be solved by technical benchmarks; we need a balanced approach with technical benchmarks, real-life testing and qualitative judgment.

UR - http://www.scopus.com/inward/record.url?scp=85064614002&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-14126-4_9

DO - 10.1007/978-3-030-14126-4_9

M3 - Chapter

SN - 978-3-030-14124-0

T3 - Cognitive Systems Monographs

SP - 169

EP - 179

BT - Metrics of Sensory Motor Coordination and Integration in Robots and Animals

PB - Springer

CY - Cham

ER -

Müller VC. Measuring progress in robotics: benchmarking and the ‘measure-target confusion’. In Bonsignorio F, Messina E, del Pobil AP, Hallam J, editors, Metrics of Sensory Motor Coordination and Integration in Robots and Animals: How to Measure the Success of Bioinspired Solutions with Respect to their Natural Models, and Against More ‘Artificial’ Solutions?. Cham: Springer. 2020. p. 169-179. (Cognitive Systems Monographs). Available from, DOI: 10.1007/978-3-030-14126-4_9