Linköping Electronic Conference Proceedings
Not a member yet
1113 research outputs found
Sort by
Adding political orientation metadata to ParlaMint corpora
Get PDFParliamentary debates are an important source for political discourse research as well as research in other disciplines. The ParlaMint project aims to create comparable corpora of parliamentary debates which, through unified encoding, provide a comprehensible resource to support such research. Within these corpora, speeches are attributed to speakers, and speaker metadata, including temporal affiliations with different organizations such as parliamentary groups and political parties. This paper discusses the addition of metadata on the political orientation of parties and parliamentary groups to the ParlaMint corpora. The paper explains our two sources for this information, namely the Chapel Hill Expert Survey Dataset and Wikipedia, the process of data collection and its subsequent encoding in the corpora. Furthermore, the paper presents an analysis of the extent of the added metadata, along with an example of exploratory data analysis. It also outlines the distribution of utterances across political orientation categories within ParlaMint, offering a comprehensive overview of the diverse perspectives and ideologies within the corpora. The inclusion of this supplementary metadata could prove valuable for parliamentary data research, while the methodology developed could be used to add further metadata to the ParlaMint corpora
Enhancing Indoor Temperature Forecasting through Synthetic Data in Low-Data Regime
Get PDFForecasting indoor temperatures is of paramount importance to achieve efficient control of HVAC systems. In this task, the limited data availability presents a challenge as most of the available data is acquired during standard operation where extreme scenarios and transitory regimes such as major temperature increases or decreases are de-facto excluded. Acquisition of such data requires significant energy consumption and a dedicated facility, hindering the quantity and diversity of available data. To acquire such data, we make use of such a facility referred to as the Test-cell. Cost related constraints however do not allow for continuous year-around acquisition.To address this, we investigate the efficacy of data augmentation techniques, particularly leveraging state-of-the-art AI-based methods for synthetic data generation. Inspired by practical and experimental motivations, we explore fusion strategies of real and synthetic data to improve forecasting models. This approach alleviates the need for continuously acquiring extensive time series data, especially in contexts involving repetitive heating and cooling cycles in buildings. Our evaluation methodology for synthetic data synthesis involves a dual-focused approach: firstly, we assess the performance of synthetic data generators independently, particularly focusing on SoTA AI-based methods; secondly, we measure the utility of incorporating synthetically augmented data in a subsequent downstream tasks (forecasting). In the forecasting tasks, we employ a simple model in two distinct scenarios: 1) we first examine an augmentation technique that combines real and synthetically generated data to expand the training dataset, 2) Second, we delve into utilizing synthetic data to tackle dataset imbalances. Our results highlight the potential of synthetic data augmentation in enhancing forecasting accuracy while mitigating training variance. Through empirical experiments, we show significant improvements achievable by integrating synthetic data, thereby paving the way for more robust forecasting models in low-data regime
Curating a historical source corpus of 20th century patient organization periodicals
Get PDFActing out Disease: How Patient Organizations Shaped Modern Medicine (ActDisease) explores the history of patient organizations in 20th century Europe. By combining traditional historiographic methods with text mining techniques, the project aims to shed light on how patient organizations co-constructed concepts of and management of disease. Part of the project is to digitize print sources and build a digital corpus for historical text mining. The corpus consists of periodical publications from selected British, French, German and Swedish patient organizations, a type of material that poses a number of challenges in scan quality, layout, and lack of consistency. This paper discusses the technical process of building the ActDisease corpus from digitizing patient organization periodicals to OCR post-processing. It touches upon the methodological questions and challenges of curating a corpus of fragmented and heterogeneous historical source material tailored to a specific project
On two SweLL learner corpora – SweLL-pilot and SweLL-gold
Get PDFSweLL – Swedish Learner Language – is a unifying term for the infrastructure module for research on Swedish as a Second Language (L2), deployed and maintained as a part of bigger infrastructure of Språkbanken Text at the University of Gothenburg, Sweden. The SweLL infrastructure module consists of a number of learner data collections, and tools for annotation and management of learner data. As a result, many of its components contain the prefix SweLL in their names, which has created some confusion, especially with regards to the two corpora. In this article we shortly introduce the various SweLL-components with a special focus on the differences between the two SweLL corpora
From the Arctics to Antarctica - A multimodular visualisation of data
Get PDFThis paper outlines the structure of Multimodal Map, a tool developed at GRIDH to access and visualise place-based datasets. The Multimodal Map frontend, which is developed with a Vue3 framework that fetches data from a backend built in Django, is arranged as a series of distinct and interconnected views that lets the user interact with the material at different scale and level of abstraction. To support the wide variety of formats the different projects need to handle, Multimodal Map makes use of both custom solutions and several open frameworks and libraries. These include Open Layers for the geographical visualisations, OpenSeadragon for IIIF-images, potree.js for point clouds, 3D Heritage Online Presenter (3DHOP) for meshes, and relight-viewer.js for RTI Photography
Standards Information System for CLARIN centres and beyond
Get PDFThe present contribution describes features of the CLARIN Standards Information System that have been designed to assist data deposition centres in CLARIN. We also show what is needed and what has been done in order to go beyond the originally designated target, so as to provide service to sibling and descendant research infrastructures, of which DARIAH and Text+ are taken as examples. This paper is aimed primarily at representatives of research infrastructure nodes, responsible for preparing and sharing data deposition information about their centres or repositories. It assumes a degree of technical knowledge or experience in using the XML format and tools, the REST API, and version control systems
Protective Measures for Sharing the Finnish DarkWeb Marketplace Corpus (FINDarC)
Get PDFWe discuss the archiving procedure of a corpus comprising posts submitted to Torilauta, a Finnish dark web marketplace website. The site was active from 2017 to 2021 and during this time one of the most prominent online illegal narcotics markets in Finland. A reduced version of the corpus, Finnish Dark Web Marketplace Corpus (FINDarC), has been archived in the Language Bank of Finland. In the current work, we focus on the protective measures for storing the data and how researchers can apply for access rights to the corpus under the CLARIN RES licence
Generating Contexts for ESP Vocabulary Exercises with LLMs
Get PDFThe current paper addresses the need for language students and teachers to have access to a large number of pedagogically sound contexts for vocabulary acquisition and testing. We investigate the automatic derivation of contexts for a vocabulary list of English for Specific Purposes (ESP). The contexts are generated by contemporary Large Language Models (namely, Mistral-7B-Instruct and Gemini 1.0 Pro) in zero-shot and few-shot settings, or retrieved from a web-crawled repository of domain-relevant websites. The resulting contexts are compared to a professionally crafted reference corpus based on their textual characteristics (length, morphosyntactic, lexico-semantic, and discourse-related). In addition, we annotated the automatically derived contexts regarding their direct applicability, comprehensibility, and domain relevance. The 'Gemini, zero-shot' contexts are rated most highly by human annotators in terms of pedagogical usability, while the 'Mistral, few-shot' contexts are globally closest to the reference based on textual characteristics
Semantic Error Prediction: Estimating Word Production Complexity
Get PDFEstimating word complexity is a well-established task in computer-assisted language learning. So far, however, complexity estimation has been largely limited to comprehension. This neglects words that are easy to comprehend, but hard to produce. We introduce semantic error prediction (SEP) as a novel task that assesses the production complexity of content words. Given the corrected version of a learner-produced text, a system has to predict which content words replace tokens from the original text. We present and analyse one example of such a semantic error prediction dataset, which we generate from an error correction dataset. As neural baselines, we use BERT, RoBERTa, and LLAMA2 embeddings for SEP. We show that our models can already improve downstream applications, such as predicting essay vocabulary scores
Private Sensitive Content on Social Media: An Analysis and Automated Detection for Norwegian
Get PDFThis study addresses the notable gap in research on detecting private-sensitive content within Norwegian social media by creating and annotating a dataset, tailored specifically to capture the linguistic and cultural nuances of Norwegian social media discourse. Utilizing Reddit as a primary data source, entries were compiled and cleaned, resulting in a comprehensive dataset of 4482 rows. Our research methodology encompassed evaluating a variety of computational models—including machine learning, deep learning, and transformers—to assess their effectiveness in identifying sensitive content. Among these, the NB BERT-based classifier emerged as the proficient, showcasing accuracy and F-1 score. This classifier demonstrated remarkable effectiveness, achieving an accuracy of 82.75% and an F1-score of 82.39%, underscoring its adeptness at navigating the complexities of privacy-sensitive content detection in Norwegian social media. This endeavor not only paves the way for enhanced privacy-sensitive content detection in Norwegian social media but also sets a precedent for future research in the domain, emphasizing the critical role of tailored datasets in advancing the field