We are developing a theory of probabilistic language learning in the context of robotic instruction in elementary assembly actions. We describe the process of machine learning in terms of the various events that happen on a given trial, including the crucial association of words with internal representations of their meaning. Of central importance in learning is the generalization from utterances to grammatical forms. Our system derives a comprehension grammar for a superset of a natural language from pairs of verbal stimuli like Go to the screw! and corresponding internal representations of coerced actions. For the derivation of a grammar no knowledge of the language to be learned is assumed but only knowledge of an internal language.We present grammars for English, Chinese, and German generated from a finite sample of about 500 commands that are roughly equivalent across the three languages. All of the three grammars, which are context-free in form, accept an infinite set of commands in the given language.
[1]
P. Suppes.
Language for humans and robots
,
1991
.
[2]
Pat Langley,et al.
Language Acquisition and Machine Learning.
,
1986
.
[3]
Jeffrey Mark Siskind,et al.
Naive physics, event perception, lexical semantics, and language acquisition
,
1992
.
[4]
John R. Anderson,et al.
Induction of Augmented Transition Networks
,
1977,
Cogn. Sci..
[5]
Patrick Suppes,et al.
Complexity Issues in Robotic Machine Learning of Natural Language
,
1992
.
[6]
Patrick Suppes,et al.
Congruence of Meaning.
,
1972
.
[7]
Andreas Stolcke.
Learning Feature-based Semantics with Simple Recurrent Networks
,
1990
.
[8]
John R. Anderson,et al.
A Theory of Language Acquisition Based on General Learning Principles
,
1981,
IJCAI.