Dataset Information

Generalising semantic category disambiguation with large lexical resources for fun and profit.

ABSTRACT:

Background

Semantic Category Disambiguation (SCD) is the task of assigning the appropriate semantic category to given spans of text from a fixed set of candidate categories, for example Protein to "Fibrin". SCD is relevant to Natural Language Processing tasks such as Named Entity Recognition, coreference resolution and coordination resolution. In this work, we study machine learning-based SCD methods using large lexical resources and approximate string matching, aiming to generalise these methods with regard to domains, lexical resources and the composition of data sets. We specifically consider the applicability of SCD for the purposes of supporting human annotators and acting as a pipeline component for other Natural Language Processing systems.

Results

While previous research has mostly cast SCD purely as a classification task, we consider a task setting that allows for multiple semantic categories to be suggested, aiming to minimise the number of suggestions while maintaining high recall. We argue that this setting reflects aspects which are essential for both a pipeline component and when supporting human annotators. We introduce an SCD method based on a recently introduced machine learning-based system and evaluate it on 15 corpora covering biomedical, clinical and newswire texts and ranging in the number of semantic categories from 2 to 91. With appropriate settings, our system maintains an average recall of 99% while reducing the number of candidate semantic categories on average by 65% over all data sets.

Conclusions

Machine learning-based SCD using large lexical resources and approximate string matching is sensitive to the selection and granularity of lexical resources, but generalises well to a wide range of text domains and data sets given appropriate resources and parameter settings. By substantially reducing the number of candidate categories while only very rarely excluding the correct one, our method is shown to be applicable to manual annotation support tasks and use as a high-recall component in text processing pipelines. The introduced system and all related resources are freely available for research purposes at: https://github.com/ninjin/simsem.

SUBMITTER: Stenetorp P

PROVIDER: S-EPMC4107982 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:In this paper I study semantic and pragmatic properties of elevational demonstratives by means of a typological investigation of 50 languages with elevational demonstratives from all across the globe. The four basic verticality values expressed by elevational demonstratives are UP, DOWN, LEVEL, and ACROSS. They can be ordered along the elevational hierarchy (UP/DOWN > LEVEL/ACROSS), which reflects cross-linguistic tendencies in the expression of these values by demonstratives. Elevational values are frequently co-expressed with distance-based meanings of demonstratives, and it is almost always distal demonstratives that express elevation, whereas medial or proximal demonstratives can lack elevational distinctions. This means that elevational demonstratives largely refer to areas outside the peripersonal sphere in a similar way as simple distal demonstratives. In the proximal domain, fine grained semantic distinctions such as those encoded by elevational demonstratives are superfluous since this domain is accessible to the interlocutors who in the default case of a normal conversation are located in close proximity to each other. I then discuss metaphorical extensions of elevational demonstratives to non-spatial uses such as temporal and social deixis. There are a few languages in which elevational demonstratives with the meaning UP express the temporal meaning future, whereas the DOWN demonstratives encode past. This finding is particularly interesting in view of the widely-debated use of Mandarin Chinese spatial terms 'up' for past events and 'down' for future events, which show the opposite metaphorical extension. I finally examine areal tendencies and potential correlations between elevational demonstratives and the geographical location of speech communities in mountainous areas such as the Himalayas, the Papuan Highlands and the Caucasus. I tentatively conclude that languages spoken in similar topographic environments do not tend to have similar systems of elevational demonstratives if they belong to different language families.