Social and Behavioral Sciences Commons^™

Learning In Minimalism-Based Language Modeling, Deryle W. Lonsdale

Faculty Publications

The natural language version of the Soar cognitive modeling system (Newell, 1990) has enabled a number of language modeling applications from on-line parsing behavior (Lewis, 1993) to simultaneous interpretation (Lonsdale, 1997, 1998) to robotic control (Benjamin, Lonsdale, & Lyons, 2004). The system supports an integrated approach to incremental comprehension and generation. Learning mechanisms account for processes in language performance from deliberate, explicit reasoning to automatic, recognitional expertise.

Syntactic processing in prior versions of the system followed the Principles and Parameters approach to syntax.

An Operator-Based Account Of Semantic Processing, Deryle W. Lonsdale, C. Anton Rytting

Faculty Publications

This paper explores issues of psychological plausibility in modeling natural language understanding within Soar, a symbolic cognitive model. It focuses on constructing syntactic and semantic representations in simulated real time, with particular emphasis on word sense disambiguation (WSD). We discuss (i) what level of WSD should be modeled and (ii) how to use resources such as WordNet to inform these models. A preliminary model of coarse-grained WSD is included to show how syntactic, semantic, and other knowledge sources interact in Soar. Finally, we explore issues of interleaving, learning, and integrating other WSD approaches with Soar's native model of learning.

Unifying Language Modeling Capabilities For Flexible Interaction, Deryle W. Lonsdale, Rebecca Diane Madsen

Faculty Publications

Dialogue processing has taken several different forms in recent years; in this paper we address a cognitive modeling approach to the problem. We begin by sketching some of our work in this area and how it derives from prior research in cognition, modeling, natural language processing, and discourse recipe compilation. We then note that, whereas the complexities of natural language— especially spoken language—have been primarily addressed by technologies that target lower-level linguistic processing, higherlevel processing has been less well studied. We introduce our approach, which is to use an agent-based cognitive modeling architecture for dialogue processing. We mention some of …

Learning In Minimalism-Based Language Modeling, Deryle W. Lonsdale

Faculty Publications

The natural language version of the Soar cognitive modeling system (Newell, 1990) has enabled a number of language modeling applications from on-line parsing behavior (Lewis, 1993) to simultaneous interpretation (Lonsdale, 1997, 1998) to robotic control (Benjamin, Lonsdale, & Lyons, 2004). The system supports an integrated approach to incremental comprehension and generation. Learning mechanisms account for processes in language performance from deliberate, explicit reasoning to automatic, recognitional expertise.

Syntactic processing in prior versions of the system followed the Principles and Parameters approach to syntax.