Text generation: using discourse strategies and focus constraints to generate natural language text
Text generation: using discourse strategies and focus constraints to generate natural language text
Conceptual structures: information processing in mind and machine
Conceptual structures: information processing in mind and machine
Conceptual graphs for the analysis and generation of sentences
IBM Journal of Research and Development
SAIRVO: a temporal planning system
IEA/AIE '89 Proceedings of the 2nd international conference on Industrial and engineering applications of artificial intelligence and expert systems - Volume 1
Maintaining knowledge about temporal intervals
Communications of the ACM
Generating Natural Language under Pragmatic Constraints
Generating Natural Language under Pragmatic Constraints
| Hi-index | 0.00 |
In project GENTEXT we aim at developing a system which will be able to generate texts in the French language, starting from knowledge structures expressed in the form of extended conceptual graphs. We had to extend the conceptual graph model mainly because in Sowa's approach temporal knowledge is not specified in a uniform way, and, monadic relations (i.e. PAST, FUTR) are not precise enough for expressing verb tenses in roman languages such as French, Spanish or Italian.In our model we use conceptual graphs for describing non-temporal knowledge structures which are called “states”. Temporal knowledge is expressed by a temporal label which is related to a state by a “period” relation (PER).In order to specify verb tenses we use an approach which was first proposed by Reichenbach and distinguishes for every statement three temporal markers: S (time of locutor), E (time at which occurred the event) and R (time of the temporal reference).This model is used for specifying conceptual structures from which sentences are generated in French.