This paper reports the state of development of PThomas , a network based document retrieval system implemented on a massively parallel fine-grained computer, the Connection Machine. The program is written in C ∗ , an enhancement of the C programming language which exploits the parallelism of the Connection Machine. The system is based on Oddy's original Thomas program, which was highly parallel in concept, and makes use of the Connection Machine's single instruction multiple data (SIMD) processing capabilities. After an introduction to systems like Thomas , and their relationship to spreading activation and neural network models, the current state of PThomas is described, including details about the network representation and the parallel operations that are executed during a typical PThomas session.

/pdf/pthomas-an-adaptive-information-retrieval-system-on-the-f95y7hwtqa.pdf

PThomas: an adaptive information retrieval system on the connection machine

Un programme de recherche «Thomas» est decrit: il cherche a repondre aux besoins de l'utilisateur en rapporchant son probleme du contenu d'une base de donnee. La strategie de recherche est basee sur un dialogue homme/machine ou le programme joue le role d'un specialiste. Une amelioration de ce programme, est presentee: Thomas II comporte une ponderation basee sur l'importance de l'association entre les documents contenus dans la base. Les experiences montrent que ce nouveau programme est plus facile a utiliser que le precedent. (INTD)

Using a cognitive model of dialogue for reference retrieval

Elsewhere, there is a description of a computer program design which aims at satisfying incompletely defined user needs through a man-machine dialogue which does not require a query formulation by the user. The program attempts to satisfy the user’s needs from a model which it builds of his interests a model which is a subset of its own ’world model’, the data base. This paper discusses current work based on this program design. Possibilities of the program creating virtual world models to prevent the user unnecessarily accessing the whole data base, as well as incorporating a form of weighting based on association strenghs between items in the data base are explored.

G.O. Ofori-Dwumfuo

Papers

Reference retrieval without user query formulation