NLI: a robust interface for natural language person-machine communication
✍ Scribed by Giovanni Guida; Carlo Tasso
- Publisher
- Elsevier Science
- Year
- 1982
- Weight
- 997 KB
- Volume
- 17
- Category
- Article
- ISSN
- 0020-7373
No coin nor oath required. For personal study only.
✦ Synopsis
In this paper, a novel approach to natural language understanding, called goal oriented parsing, is presented. Such a model of comprehension is embedded in the more general frame of interpersonal communication and is applied to the person-machine interaction task. It is based on the claim that understanding imperative natural language is strongly dependent both on the goal of the speaker and on the nature of the hearer. This assumption has proved appropriate for the design of effective and robust natural language interfaces to artificial systems. This approach is supported by the development of an experimental project, called NLI, for enquiring in Italian a relational data base. NLI is to date running on a PDP 11/34 computer at the University of Udine. In this paper we illustrate the overall architecture of the system, along with the basic features of the parsing algorithm. This is based on the new concept of hierarchical parsing and is mainly directed by the semantics of the language. The role of clarification dialogue to overcome critical situations is discussed as well. The use of goal oriented parsing in dealing with full query text and anaphora is also exploited. Several meaningful examples are presented. Comparisons with related works are outlined and promising directions deserving further investigation are presented.