A Hybrid and Connectionist Architecture for a Scanning Understanding
نویسنده
چکیده
This paper describes a general architecture SCAN for hybrid symbolic connectionist processing of natural language phrases. SCAN's architecture shows how learned connectionist domain-dependent semantic representations can be combined with encoded symbolic syntactic representations. Within this general architecture we focus on a connectionist model for semantic classiication based on a scanning understanding of phrases. We specify strategies at the topmost theory level and we show how these strategies are realized in a recurrent connectionist plausibility network at the underlying representation level. In particular, this model demonstrates that a recurrent connectionist network can learn a semantic memory model for phrase classiication based on a scanning understanding.
منابع مشابه
Spoken Language Processing in the Hybrid Connectionist Architecture SCREEN
In this paper1 we describe a robust, learning approach to spoken language understanding. Since interactively spoken and computationally analyzed language often contains many errors, robust connectionist networks are used for providing a at screening analysis. A screening analysis is a shallow at analysis based on category sequences at various syntactic, semantic and dialog levels. Rather than u...
متن کاملA Hybrid Connectionist and BDI Architecture for Modeling Embedded Rational Agents
In this paper, our ongoing work on a hybrid, connectionist and belief-desire-intention (BDI) based, ratiohal agent architecture is described. The architecture makes specific commitments in order to achieve a harmony among the tasks of reasoning and acting. The architecture uses a bottom-up learning strategy to acquire rules for reactive behavior. Embedded Agents: From Reactive to
متن کاملThe DUAL Cognitive Architecture: A Hybrid Multi-Agent Approach
A hybrid (symbolic/connectionist) cognitive architecture, DUAL, is proposed. It is a multi-agent system which consist of a large number of non-cognitive, relatively simple agents, and which behaviour emerges from the behaviour of these simple agents and the interactions between them. The agents within this architecture have no internal knowledge base and goals. They are both computational devic...
متن کاملHybrid Approaches to Neural Network-based Language Processing
In this paper we outline hybrid approaches to arti cial neural network-based natural language processing. We start by motivating hybrid symbolic/connectionist processing. Then we suggest various types of symbolic/connectionist integration for language processing: connectionist structure architectures, hybrid transfer architectures, hybrid processing architectures. Furthermore, we focus particul...
متن کاملA Hybrid Learning Model of Abductive Reasoning
Multicausal abductive tasks appear to have deliberate and implicit components: people generate and modify explanations using a series of recognizable steps, but these steps appear to be guided by an implicit hypothesis evaluation process. This paper proposes a hybrid symbolic-connectionist learning architecture for multicausal abduction. The architecture tightly integrates a symbolic Soar model...
متن کامل