Information-theoretic measures are among the most standard techniques for evaluation of clustering methods including word sense induction (WSI) systems. Such measures rely on sample-based estimates of the entropy. However, the standard maximum likelihood estimates of the entropy are heavily biased with the bias dependent on, among other things, the number of clusters and the sample size. This makes the measures unreliable and unfair when the number of clusters produced by different systems vary and the sample size is not exceedingly large. This corresponds exactly to the setting of WSI evaluation where a ground-truth cluster sense number arguably does not exist and the standard evaluation scenarios use a small number of instances of each word to compute the score. We describe more accurate entropy estimators and analyze their performance both in simulations and on evaluation of WSI systems.

Traditionally, most research in NLP has focused on propositional aspects of meaning. To truly understand language, however, extra-propositional aspects are equally important. Modality and negation typically contribute significantly to these extra-propositional meaning aspects. Although modality and negation have often been neglected by mainstream computational linguistics, interest has grown in recent years, as evidenced by several annotation projects dedicated to these phenomena. Researchers have started to work on modeling factuality, belief and certainty, detecting speculative sentences and hedging, identifying contradictions, and determining the scope of expressions of modality and negation. In this article, we will provide an overview of how modality and negation have been modeled in computational linguistics.

Abstract Being able to identify which rhetorical relations (e.g., contrast or explanation) hold between spans of text is important for many natural language processing applications. Using machine learning to obtain a classifier which can distinguish between different relations typically depends on the availability of manually labelled training data, which is very time-consuming to create. However, rhetorical relations are sometimes lexically marked, i.e., signalled by discourse markers (e.g., because, but, consequently etc.), and it has been suggested (Marcu and Echihabi, 2002) that the presence of these cues in some examples can be exploited to label them automatically with the corresponding relation. The discourse markers are then removed and the automatically labelled data are used to train a classifier to determine relations even when no discourse marker is present (based on other linguistic cues such as word co-occurrences). In this paper, we investigate empirically how feasible this approach is. In particular, we test whether automatically labelled, lexically marked examples are really suitable training material for classifiers that are then applied to unmarked examples. Our results suggest that training on this type of data may not be such a good strategy, as models trained in this way do not seem to generalise very well to unmarked data. Furthermore, we found some evidence that this behaviour is largely independent of the classifiers used and seems to lie in the data itself (e.g., marked and unmarked examples may be too dissimilar linguistically and removing unambiguous markers in the automatic labelling process may lead to a meaning shift in the examples).