Automatic Generation of On-Line Documentation in the IDAS Project
40906Authors
nan
Abstract
The Intelligent Documentation Advisory Sys- tem generates on-line documentation and help messages from a domain knowledge base, using natural-language (NL) generation techniques. This paper gives an overview of IDAS, with par- ticular emphasis on: (1) its architecture and the types of questions it is capable of answering; (2) its KR and NL generation systems, and lessons we have learned in designing them; and (3) its hypertext-like user interface, and the benefits such an interface brings. 1
Keywords
Full Content
Automatic Generation of On-Line Documentation in the IDAS Project* Ehud Reiter, Chris Mellish, and John Levine Department of Artificial Intelligence University of Edinburgh 80 South Bridge Edinburgh EH1 1HN BRITAIN (e-mail: e.reiter@ed.ac.uk) Abstract The Intelligent Documentation Advisory Sys- tem generates on-line documentation and help messages from a domain knowledge base, using natural-language (NL) generation techniques. This paper gives an overview of IDAS, with par- ticular emphasis on: (1) its architecture and the types of questions it is capable of answering; (2) its KR and NL generation systems, and lessons we have learned in designing them; and (3) its hypertext-like user interface, and the benefits such an interface brings. 1 Introduction The Intelligent Documentation Advisory System (IDAS) project is attempting to use natural-language (NL) gen- eration and hypertext technology to produce an on-line documentation and help system that supports users of complex machinery. In this paper we present an overview of the most recent IDAS prototype developed at the Uni- versity of Edinburgh, including descriptions of: • IDAS's overall architecture, and particularly its queslion space, i.e., the set of queries it is designed to answer; • IDAS's I(R and NL generation components, with particular emphasis on lessons we have learned while building them, and related design decisions; • IDAS's hypertext-like user interface, and the oper- ations it is intended to support. IDAS is a collaborative effort between the University of Edinburgh, Raeal Instruments Ltd., Racal Research • The IDAS project is partially funded by UK SERC grant GR/F/36750 and UI( DTI grant IED 4/1/1072, and we are grateful to SERC and DTI for their support of this work. We would also like to thank the IDAS industrial collaborators -- Inference Europe, Ltd.; Racal Instruments, Ltd.; and Racal Research, Ltd. -- for all the help they have given us in performing this research. Thanks also to Robert Dale and tile anonymous reviewers for their very helpful comments Ltd., and Inference Europe Ltd. As this paper is written, the project is about half-way through its 3-year lifespan. Several prototypes have been built to date; this paper describes the most recent one built at the University of Edinburgh, which consists of about 5000 lines of Lisp code. Work on IDAS continues as this paper is being written, with current tasks including the expansion of the existing domain knowledge base, and the integration of the documentation software with the actual hardware being documented. We have not yet carried out any formal evaluations of IDAS, although we hope to arrange such tests once the current expansion and integration tasks are completed; the existing system has been shown to many people informally, generally with quite favorable reactions. The initial IDAS system documents an ATE (Auto- matic Test Equipment), a complex device made by Racal Instruments for testing potentially faulty electronic de- vices. The ATE contains an assortment of electronic instruments, a switching system that connects these in- struments to a UUT (Unit Under Test), and a computer which runs test programs that test the UUT with the …
Content truncated to 500 words
Labeled Text
The Intelligent Documentation Advisory Sys- tem generates on-line documentation and help messages from a domain knowledge base, using natural-language (NL) generation techniques. This paper gives an overview of IDAS, with par- ticular emphasis on: (1) its architecture and the types of questions it is capable of answering; (2) its KR and NL generation systems, and lessons we have learned in designing them; and (3) its hypertext-like user interface, and the benefits such an interface brings. 1
Publication Details
- Paper ID:
- 40906
- Year:
- 1992
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan