HMM-based concept learning for a mobile robot

Kevin M. Squire, Stephen E. Levinson

Research output: Contribution to journalArticle

Abstract

We are developing an intelligent robot and attempting to teach it language. While there are many aspects of this research, for the purposes here the most important are the following ideas. Language is primarily based on semantics, not syntax, which is still the focus in speech recognition research these days. To truly learn meaning, a language engine cannot simply be a computer program running on a desktop computer analyzing speech. It must be part of a more general, embodied intelligent system, one capable of using associative learning to form concepts from the perception of experiences in the world, and further capable of manipulating those concepts symbolically. In this paper, we present a general cascade model for learning concepts, and explore the use of hidden Markov models (HMMs) as part of the cascade model. HMMs are capable of automatically learning and extracting the underlying structure of continuous-valued inputs and representing that structure in the states of the model. These states can then be treated as symbolic representations of the inputs. We show how a cascade of HMMs can be embedded in a small mobile robot and used to find correlations among sensory inputs to learn a set of symbolic concepts, which are used for decision making and could eventually be manipulated linguistically.

Original languageEnglish (US)
Pages (from-to)199-212
Number of pages14
JournalIEEE Transactions on Evolutionary Computation
Volume11
Issue number2
DOIs
StatePublished - Apr 1 2007

Fingerprint

Concept Learning
Hidden Markov models
Mobile Robot
Mobile robots
Markov Model
Cascade
Model-based
Intelligent robots
Intelligent systems
Intelligent Systems
Speech Recognition
Speech recognition
Personal computers
Computer program listings
Engine
Robot
Decision making
Decision Making
Semantics
Model

Keywords

  • Developmental robotics
  • Hidden Markov models (HMMs)
  • Hierarchical model
  • Online learning
  • Semantic learning

ASJC Scopus subject areas

  • Software
  • Theoretical Computer Science
  • Computational Theory and Mathematics

Cite this

HMM-based concept learning for a mobile robot. / Squire, Kevin M.; Levinson, Stephen E.

In: IEEE Transactions on Evolutionary Computation, Vol. 11, No. 2, 01.04.2007, p. 199-212.

Research output: Contribution to journalArticle

@article{8814c9d815f1407a8505d124eaeeb4c8,
title = "HMM-based concept learning for a mobile robot",
abstract = "We are developing an intelligent robot and attempting to teach it language. While there are many aspects of this research, for the purposes here the most important are the following ideas. Language is primarily based on semantics, not syntax, which is still the focus in speech recognition research these days. To truly learn meaning, a language engine cannot simply be a computer program running on a desktop computer analyzing speech. It must be part of a more general, embodied intelligent system, one capable of using associative learning to form concepts from the perception of experiences in the world, and further capable of manipulating those concepts symbolically. In this paper, we present a general cascade model for learning concepts, and explore the use of hidden Markov models (HMMs) as part of the cascade model. HMMs are capable of automatically learning and extracting the underlying structure of continuous-valued inputs and representing that structure in the states of the model. These states can then be treated as symbolic representations of the inputs. We show how a cascade of HMMs can be embedded in a small mobile robot and used to find correlations among sensory inputs to learn a set of symbolic concepts, which are used for decision making and could eventually be manipulated linguistically.",
keywords = "Developmental robotics, Hidden Markov models (HMMs), Hierarchical model, Online learning, Semantic learning",
author = "Squire, {Kevin M.} and Levinson, {Stephen E.}",
year = "2007",
month = "4",
day = "1",
doi = "10.1109/TEVC.2006.890263",
language = "English (US)",
volume = "11",
pages = "199--212",
journal = "IEEE Transactions on Evolutionary Computation",
issn = "1089-778X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

TY - JOUR

T1 - HMM-based concept learning for a mobile robot

AU - Squire, Kevin M.

AU - Levinson, Stephen E.

PY - 2007/4/1

Y1 - 2007/4/1

N2 - We are developing an intelligent robot and attempting to teach it language. While there are many aspects of this research, for the purposes here the most important are the following ideas. Language is primarily based on semantics, not syntax, which is still the focus in speech recognition research these days. To truly learn meaning, a language engine cannot simply be a computer program running on a desktop computer analyzing speech. It must be part of a more general, embodied intelligent system, one capable of using associative learning to form concepts from the perception of experiences in the world, and further capable of manipulating those concepts symbolically. In this paper, we present a general cascade model for learning concepts, and explore the use of hidden Markov models (HMMs) as part of the cascade model. HMMs are capable of automatically learning and extracting the underlying structure of continuous-valued inputs and representing that structure in the states of the model. These states can then be treated as symbolic representations of the inputs. We show how a cascade of HMMs can be embedded in a small mobile robot and used to find correlations among sensory inputs to learn a set of symbolic concepts, which are used for decision making and could eventually be manipulated linguistically.

AB - We are developing an intelligent robot and attempting to teach it language. While there are many aspects of this research, for the purposes here the most important are the following ideas. Language is primarily based on semantics, not syntax, which is still the focus in speech recognition research these days. To truly learn meaning, a language engine cannot simply be a computer program running on a desktop computer analyzing speech. It must be part of a more general, embodied intelligent system, one capable of using associative learning to form concepts from the perception of experiences in the world, and further capable of manipulating those concepts symbolically. In this paper, we present a general cascade model for learning concepts, and explore the use of hidden Markov models (HMMs) as part of the cascade model. HMMs are capable of automatically learning and extracting the underlying structure of continuous-valued inputs and representing that structure in the states of the model. These states can then be treated as symbolic representations of the inputs. We show how a cascade of HMMs can be embedded in a small mobile robot and used to find correlations among sensory inputs to learn a set of symbolic concepts, which are used for decision making and could eventually be manipulated linguistically.

KW - Developmental robotics

KW - Hidden Markov models (HMMs)

KW - Hierarchical model

KW - Online learning

KW - Semantic learning

UR - http://www.scopus.com/inward/record.url?scp=34047269122&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34047269122&partnerID=8YFLogxK

U2 - 10.1109/TEVC.2006.890263

DO - 10.1109/TEVC.2006.890263

M3 - Article

AN - SCOPUS:34047269122

VL - 11

SP - 199

EP - 212

JO - IEEE Transactions on Evolutionary Computation

JF - IEEE Transactions on Evolutionary Computation

SN - 1089-778X

IS - 2

ER -