NALUPES – Natural Language Understanding and Processing Expert System

Abstract

Today, modern electronic devices are supplied with many new sophisticated functions, and expectations of their users are constantly growing. Abilities of natural language handling, i.e. understanding and processing commands given in natural language undoubtedly increase the attraction of such equipment. Moreover, this property makes them useful for those persons who have problems with standard communication, with limited manual dexterity, handicapped or even blind persons. This problem can be extended to other fields of human’s activity. Also electronic design automation (EDA vendors work on facilitation of the design process for engineers and simplification of their tools. (Pulka & Klosowski, 2009) described the idea of such expert system supplied with the speech recognition module, dialog module with speech synthesis elements and inference engine responsible for data processing and language interpreting. That approach is dedicated to system-level electronic design problems. The authors focused on automatic generation of modules based on speech and language processing and on data manipulating. This chapter focuses on highest level language processing phase the heart of the system – the intelligent expert system responsible for appropriate interpretation of commands given in natural language and formulation of responses to the user. We concentrate on inference engine that works on the text strings that are far from the lowest, signal level. Automated processing and understanding of natural language have been recognized for years and we can find these problems in many practical applications (Manning & Schultze, 1999, Jurafsky & Martin, 2000). They belong to hot topics investigated in many academic centers (Gu et al. 2006, Ammicht et al. 2007, Infantino et al. 2007, Neumeier & Thompson 2007, Wang 2007). The main objective of the presented contribution is to develop an expert system that aids the design process and enriches its abilities with speech recognition and speech synthesis properties. The proposed solution is intended to be an optional tool incorporated into the more complex environment working in the background. The goal is to create a system that assists the working. These objectives can be met with the AI-based expert system consisting of the following components: speech recognition module, speech synthesis module, language processing module with knowledge base (dictionary and semantic rules), knowledge base of the design components and design rules and the intelligent inference engine which ties together entire system, controls the data traffic and checks if the user demands are correctly interpreted.