Skip to main content

Intelligent Documents Production Demonstrator

Objective

The objectives of INDOC were to:
-validate an approach incorporating the use of AI techniques in Advanced Information Systems for the production of complex documents
-define the requirements of a user tool-set for document development and maintenance.
The production of complex documents such as a contract, a licence, an insurance policy, etc, requires the merging of operational information (data and images), handled by traditional tools, with advanced text processing.
This project investigated AI techniques that were thought capable of contributing to this process by developing and maintaining large sets of conceptual descriptions of documents and organizational procedures.
The feasibility of the approach was to be tested on a restricted but real-life application domain, for which a demonstrator was to be built. An evaluation of the results was designated to assess the portability of the approach to a large application.
Te objectives were to validate an approach incorporating the use of artificial intelligence (AI) techniques in advanced information systems for the production of complex documents, and to define the requirements of a user tool set for document development and maintenance. The production of complex documents requires the merging of operational information (dataand images), handled by traditional tools, with advanced text processing. This project investigated AI techniques that were thought capable of contributing to this process by developing and maintaining large sets of conceptual descriptions of documents and organizational procedures. The feasibility of the approach was tested on a restricted but real life application domain, for which a demonstrator was built. An evaluation of the results was designated to assess the portability of the approach to a large application. A methodology for knowledge acquisition and analysis which provides a reference pattern for questioning experts and classifying the knowledge acquired was agreed. The application domain for the demonstrator was the creation of legal deeds and contracts. The 2 cases analysed were the production of deeds of gift and sale by a lawyer's office, and the production of deeds for a loan receipt by a bank. A formalized behavioural model was defined, and the requirements for a formal language for conceptual modelling stated. A language for the definition of presentation schema was specified. A prototype of the inference machine is being implemented in Prolog.
A methodology for knowledge acquisition and analysis which provides a reference pattern for questioning experts and classifying the knowledge acquired was agreed.
The application domain for the demonstrator was the creation of legal deeds and contracts. The two cases analysed were the production of deeds of gift and sale by a lawyer's office, and the production of deeds for a loan receipt by a bank.
A formalised behavioural model was defined, and the requirements for a formal language for conceptual modelling stated. A language for the definition of presentation schema was specified. A prototype of the inference machine is being implemented in Prolog.
Exploitation
The INDOC project made a strong contribution to the fundamental harmonisation of AI techniques by proving and verifying their effectiveness in a real-life application. The demonstrator produced by INDOC can be exploited by European industry to evaluate the use of AI and KB techniques within real-life information systems.

Coordinator

Applied Research Group SpA
Address
Via Pio La Torre 14
20090 Vimodrone Milano
Italy

Participants (2)

EPSILON SOFTWARE LTD
Greece
Address
Syngrou Avenue 190
17671 Athinai
INESC-INSTITUTO DE ENGENHARIA DE SISTEMAS E COMPUTADORES
Portugal
Address
Apartado 10105, Rua Alves Redol
1017 Lisboa