Multilingual generation of uncertain temporal expressions from data: A study of a possibilistic formalism and its consistency with human subjective evaluations

Abstract

In NLG systems, temporal uncertainty in raw data can hamper the inference of temporal and causal relationships between events and thus impact the quality of the generated texts. In this paper, we introduce a framework to represent and reason with temporal uncertainty based on possibility theory and propose a model that uses the outcomes of such temporal reasoning to select linguistic expressions to convey uncertainty to the reader. Our model is based on Fuzzy Temporal Constraint Networks (FTCN) and our work is based on the assumption that uncertainty should be communicated to an end user. The model we propose is grounded in experimental data from three languages. We present a large-scale empirical study that investigates the conditions that influence human subjective uncertainty in reasoning about temporal relations. Based on this, we also construct a classifier to select expressions to convey uncertainty, based on possibility and necessity values. We then present an evaluation which shows that the predictions of the FTCN model correlate well with human subjective uncertainty in different scenarios. An evaluation of our temporal expressions classifier also suggests good results, compared to human selection of linguistic expressions, as compared to baseline models.

The Role of Graduality for Referring Expression Generation in Visual Scenes

“…Many techniques have been described to handle fuzzy spatial relationships, and we believe that reg would benefit from an exploration of this literature. Some related work has shown that fuzzy sets can be useful in generating appropriate natural language descriptions of uncertain temporal relations between events [48]; applying similar strategies to spatial relations is a promising way forward.…”

Fuzzy sets across the natural language generation pipeline

“…Alternatively, automatically learn such definitions (for instance, as it is done in TSK rule-based systems [19]). • Perform experiments to obtain empirical definitions from experts or human users [14], [20].…”

“…In other cases experiments were run in order to study how human subjects use linguistic expressions in different domains (e.g. [14], [20]).…”

“…Likewise, NLG systems may use LDD and other fuzzy-related techniques to address the problem of uncertainty at different levels within the distinct tasks involved in an NLG process. Examples from both trends include the GALiWeather [13] system, conceived as a LDD approach but made use of several NLG techniques to generate short-term weather forecasts; and the model presented in [14] that involves fuzzy temporal constraint networks and experimental data from three languages to address the problem of generating uncertain temporal expressions from an NLG point of view.…”

How to talk to strangers: Generating medical reports for first-time users

“…There are several lines of research aiming to address this problem. For example, previous work has focused on addressing uncertainty in time-series data [17] and temporal uncertainty [29]. Our study addresses uncertainty derived from unknown first-time users.…”