The scope and limits of simulation in automated reasoning

Ernest Davis, Gary Marcus

Research output: Contribution to journalArticle

Abstract

In scientific computing and in realistic graphic animation, simulation - that is, step-by-step calculation of the complete trajectory of a physical system - is one of the most common and important modes of calculation. In this article, we address the scope and limits of the use of simulation, with respect to AI tasks that involve high-level physical reasoning. We argue that, in many cases, simulation can play at most a limited role. Simulation is most effective when the task is prediction, when complete information is available, when a reasonably high quality theory is available, and when the range of scales involved, both temporal and spatial, is not extreme. When these conditions do not hold, simulation is less effective or entirely inappropriate. We discuss twelve features of physical reasoning problems that pose challenges for simulation-based reasoning. We briefly survey alternative techniques for physical reasoning that do not rely on simulation.

Original languageEnglish (US)
Pages (from-to)60-72
Number of pages13
JournalArtificial Intelligence
Volume233
DOIs
StatePublished - Apr 1 2016

Fingerprint

Natural sciences computing
simulation
Animation
Trajectories
artificial intelligence
Automated Reasoning
Simulation
available information
Physical

Keywords

  • Physical reasoning
  • Simulation

ASJC Scopus subject areas

  • Artificial Intelligence
  • Language and Linguistics
  • Linguistics and Language

Cite this

The scope and limits of simulation in automated reasoning. / Davis, Ernest; Marcus, Gary.

In: Artificial Intelligence, Vol. 233, 01.04.2016, p. 60-72.

Research output: Contribution to journalArticle

@article{250bd9659cb043bfabb89ce93a6fdbbf,
title = "The scope and limits of simulation in automated reasoning",
abstract = "In scientific computing and in realistic graphic animation, simulation - that is, step-by-step calculation of the complete trajectory of a physical system - is one of the most common and important modes of calculation. In this article, we address the scope and limits of the use of simulation, with respect to AI tasks that involve high-level physical reasoning. We argue that, in many cases, simulation can play at most a limited role. Simulation is most effective when the task is prediction, when complete information is available, when a reasonably high quality theory is available, and when the range of scales involved, both temporal and spatial, is not extreme. When these conditions do not hold, simulation is less effective or entirely inappropriate. We discuss twelve features of physical reasoning problems that pose challenges for simulation-based reasoning. We briefly survey alternative techniques for physical reasoning that do not rely on simulation.",
keywords = "Physical reasoning, Simulation",
author = "Ernest Davis and Gary Marcus",
year = "2016",
month = "4",
day = "1",
doi = "10.1016/j.artint.2015.12.003",
language = "English (US)",
volume = "233",
pages = "60--72",
journal = "Artificial Intelligence",
issn = "0004-3702",
publisher = "Elsevier",

}

TY - JOUR

T1 - The scope and limits of simulation in automated reasoning

AU - Davis, Ernest

AU - Marcus, Gary

PY - 2016/4/1

Y1 - 2016/4/1

N2 - In scientific computing and in realistic graphic animation, simulation - that is, step-by-step calculation of the complete trajectory of a physical system - is one of the most common and important modes of calculation. In this article, we address the scope and limits of the use of simulation, with respect to AI tasks that involve high-level physical reasoning. We argue that, in many cases, simulation can play at most a limited role. Simulation is most effective when the task is prediction, when complete information is available, when a reasonably high quality theory is available, and when the range of scales involved, both temporal and spatial, is not extreme. When these conditions do not hold, simulation is less effective or entirely inappropriate. We discuss twelve features of physical reasoning problems that pose challenges for simulation-based reasoning. We briefly survey alternative techniques for physical reasoning that do not rely on simulation.

AB - In scientific computing and in realistic graphic animation, simulation - that is, step-by-step calculation of the complete trajectory of a physical system - is one of the most common and important modes of calculation. In this article, we address the scope and limits of the use of simulation, with respect to AI tasks that involve high-level physical reasoning. We argue that, in many cases, simulation can play at most a limited role. Simulation is most effective when the task is prediction, when complete information is available, when a reasonably high quality theory is available, and when the range of scales involved, both temporal and spatial, is not extreme. When these conditions do not hold, simulation is less effective or entirely inappropriate. We discuss twelve features of physical reasoning problems that pose challenges for simulation-based reasoning. We briefly survey alternative techniques for physical reasoning that do not rely on simulation.

KW - Physical reasoning

KW - Simulation

UR - http://www.scopus.com/inward/record.url?scp=84953790112&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84953790112&partnerID=8YFLogxK

U2 - 10.1016/j.artint.2015.12.003

DO - 10.1016/j.artint.2015.12.003

M3 - Article

VL - 233

SP - 60

EP - 72

JO - Artificial Intelligence

JF - Artificial Intelligence

SN - 0004-3702

ER -