Skip to main content

Perspectives

GRaSP​

Our formal framework for modeling perspectives and their sources on content is called GRaSP (Grounded Representation and Source Perspective). This paper explains how this framework provides a way to connect statements (in texts, video, images, etc.) to their source, the entities and events they mention, and the attitude of the source towards this content. This repository provides the GRaSP ontology as well as examples illustrating GRaSP.

See also the nwr-triple-api repository, for querying the KnowledgeStore populated with NewsReader output and representing the result as SEM-RDF or SEM-JSON

Multilingual Event Factuality​

We are building a multi-lingual system for identifying the factuality of events. This repository contains the first version of the vua-perspectives-factuality module. The code for a basic implementation for Dutch and English can be found in here. In addition, we provide various scripts (and some data) for running experiments on factuality classifiers here. The setup is designed for modules using NAF and evaluation data annotated with the CAT tool. The current version supports experiments on the NewsReader MEANTIME corpus.

Affixal Negation​

We have created an affixal negation dictionary by annotating all direct antonym pairs in WordNet using an existing typology of affixal negation. The work is described in this paper and the annotations can be found here.

Negation & Inference​

This repository contains the code for conducting the experiments as reported in this paper. We reimplemented a system for detecting implicit positive meaning from negated statements as described by Blanco and Sarabi (2016) and propose a simple baseline that outperforms the system.

Microportraits​

A microportrait is an impression of a target group or individual conveyed in a single text. This repository contains code to extract microportraits as described in this paper, where we use microportraits extraction for investigating stereotyping of Muslims in the Dutch media.

Storyline Extraction​

this repository contains the materials associated with the StoryLine Extraction Task. The annotation scheme is described in this paper and the corpus is described in this paper.

Biographies​

Age discrimination​