Search

Most of the work on Character Networks to date is limited to monolingual texts. Conversely, in this paper we apply and analyze Character Networks on both source texts (English novels) and their Finnish translations (both human- and machine-translated). We assume that this analysis could provide some insights on changes in translations that could modify the character networks, as well as the narrative. The results show that the character networks of translations differ from originals in case of long novels, and the differences may also vary depending on the novel and translator's strategy.

While the production of information in the European early modern period is a well-researched topic, the question how people were engaging with the information explosion that occurred in early modern Europe, is still underexposed. This paper presents the annotations and experiments aimed at exploring whether we can automatically extract media related information (source, perception, and receiver) from a corpus of early modern Dutch chronicles in order to get insight in the mediascape of early modern middle class people from a historic perspective. In a number of classification experiments with Conditional Random Fields, three categories of features are tested: (i) raw and binary word embedding features, (ii) lexicon features, and (iii) character features. Overall, the classifier that uses raw embeddings performs slightly better. However, given that the best F-scores are around 0.60, we conclude that the machine learning approach needs to be combined with a close reading approach for the results to be useful to answer history research questions.

Pretrained language models like BERT have advanced the state of the art for many NLP tasks. For resource-rich languages, one has the choice between a number of language-specific models, while multilingual models are also worth considering. These models are well known for their crosslingual performance, but have also shown competitive in-language performance on some tasks. We consider monolingual and multilingual models from the perspective of historical texts, and in particular for texts enriched with editorial notes: how do language models deal with the historical and editorial content in these texts? We present a new Named Entity Recognition dataset for Dutch based on 17th and 18th century United East India Company (VOC) reports extended with modern editorial notes. Our experiments with multilingual and Dutch pretrained language models confirm the crosslingual abilities of multilingual models while showing that all language models can leverage mixed-variant data. In particular, language models successfully incorporate notes for the prediction of entities in historical texts. We also find that multilingual models outperform monolingual models on our data, but that this superiority is linked to the task at hand: multilingual models lose their advantage when confronted with more semantical tasks.

In this article, we discuss which text, speech, and image technologies have been developed, and would be feasible to develop, for the approximately 60 Indigenous languages spoken in Canada. In particular, we concentrate on technologies that may be feasible to develop for most or all of these languages, not just those that may be feasible for the few most-resourced of these. We assess past achievements and consider future horizons for Indigenous language transliteration, text prediction, spell-checking, approximate search, machine translation, speech recognition, speaker diarization, speech synthesis, optical character recognition, and computer-aided language learning., 27th International Conference on Computational Linguistics, August 20-26, 2018, Santa Fe, New Mexico, USA

A Studied is a promising new kind of microbiotesting: the dynamics of galvanotaxis of infu-sorians. Explanations are given to this effect, as to the repeated movement of the population of infu-sorians when the polarity of the electrodes is changed. Materials showing the relevance and interest to the method by experts in the field of biotechnology, toxicology and law are presented. The question about the necessity and complexity of the mathematical model for this method is raised. Models reflecting various aspects of galvanotaxis of infusoria are considered, and the range of scientific studies from XIX to XXI centuries is covered. The conclusion is made that it is necessary to create a new model taking into account the former ones, but having a specific microbiota test based on the experiment, with the population movement of infusorians. The experiment has been carried out using a SF-56 spectrophotometer and a cuvette holder to obtain a signal near the electrode. The tasks of creating a model are formulated, and a conclusion is made about the recurrent model. The features of recurrent processes and parameters which allow using them to determine the adequacy of the model and experiment are considered. The principles of the recurrent model are used as the basis for the formulas for calculating the number of particles in the cuvette zones, which made it possible to create several types of computer programs which allowed obtaining a pulse amplitude sequence controlled by the particle transition coefficients. In recurrent simulation of the galvanotaxis dynamics process it was proved that the envelope of the amplitudes retains the function under different influences: with a control model and toxicity, which is adequate to the function obtained in the experiment. It was also verified that the initial particle distributions do not affect the correlation of the envelope function, which is typical for recurrent models. Approaches to the problem of toxicity according to the function of the envelope are proposed.

This dissertation describes a research project concerned with establishing the topical structure of long informal documents. In this research, we place special emphasis on literary data, but also work with speech transcripts and several other types of data. It has long been acknowledged that discourse is more than a sequence of sentences but, for the purposes of many Natural Language Processing tasks, it is often modelled exactly in that way. In this dissertation, we propose a practical approach to modelling discourse structure, with an emphasis on it being computationally feasible and easily applicable. Instead of following one of the many linguistic theories of discourse structure, we attempt to model the structure of a document as a tree of topical segments. Each segment encapsulates a span that concentrates on a particular topic at a certain level of granularity. Each span can be further sub-segmented based on finer fluctuations of topic. The lowest (most refined) level of segmentation is individual paragraphs. In our model, each topical segment is described by a segment centre -- a sentence or a paragraph that best captures the contents of the segment. In this manner, the segmenter effectively builds an extractive hierarchical outline of the document. In order to achieve these goals, we use the framework of factor graphs and modify a recent clustering algorithm, Affinity Propagation, to perform hierarchical segmentation instead of clustering. While it is far from being a solved problem, topical text segmentation is not uncharted territory. The methods developed so far, however, perform least well where they are most needed: on documents that lack rigid formal structure, such as speech transcripts, personal correspondence or literature. The model described in this dissertation is geared towards dealing with just such types of documents. In order to study how people create similar models of literary data, we built two corpora of topical segmentations, one flat and one hierarchical. Each document in these corpora is annotated for topical structure by 3-6 people. The corpora, the model of hierarchical segmentation and software for segmentation are the main contributions of this work.

This work is an inquiry into automatic summarization of short of fiction. In this dissertation, I present a system that composes summaries of literary short stories employing two types of information: information about entities central to a story and information about the grammatical aspect of clauses. The summaries are tailored to a specific purpose: helping a reader decide whether she would be interested in reading a particular story. They contain just enough information to enable a reader to form adequate expectations about the story, but they do not reveal the plot. According to these criteria, a target summary provides a reader with an idea of whom the story is about, where and when it happens (in a way that goes beyond simply listing names and places) but does not re-tell the events of the story. In order to build such summaries, the system attempts to identify sentences that meet two criteria: they focus on main entities in the story and they relate the background of the story rather than events. Discussing the criteria for the sentence selection process comprises a large part of this dissertation. These criteria can be roughly divided into two categories: (1) information about main entities (e.g., main characters and locations) and (2) information related to the grammatical aspect of clauses. By relying on this information the system selects sentences that contain important information pertinent to the setting of the story. Six human judges evaluated the produced summaries in two different ways. Initially, the machine-made summaries were compared against man-made ones. On this account, the summaries rated better than those produced using two naive lead-based baselines. Subsequently, the judges answered a number of questions using the summaries as the only source of information. These answers were compared with the answers made using the complete stories. The summaries appeared to be useful for helping the judges decide whether they would like to read the stories. The judges could also answer simple questions about the setting of the story using the summaries only. The results suggest that aspectual information and information about important entities can be effectively used to build summaries of literary short fiction, even though this information atone is not sufficient for producing high-quality indicative summaries.

This paper studies the use of NMT (neural machine translation) as a normalization method for an early English letter corpus. The corpus has previously been normalized so that only less frequent deviant forms are left out without normalization. This paper discusses different methods for improving the normalization of these deviant forms by using different approaches. Adding features to the training data is found to be unhelpful, but using a lexicographical resource to filter the top candidates produced by the NMT model together with lemmatization improves results.