Search

As the primary resource for preschool children, picture books,and their gender narratives, can unconsciously shape and changechildren’s perceptions of sex roles and gender identity (Bleakley etal., 1988; Connor & Serbin, 1978; Latima, 2020). However, existingstudies show concerning trends in the representation of genderinequality in modern picture books, such as the overwhelmingnumber of male main characters and traditional gender stereotypesof vocations, personalities and habits (Casey et al., 2021; Hamiltonet al., 2006; Lee & Chin, 2019; Terras, 2018). It is thereforeimportant for children’s picture books to have diverse gender descriptionsand improved equal gender representations.Since the second Feminist Movement in the 1960s, gender equalityin UK children’s picture books have been continuously examinedyet slowly improved (Adams et al., 2011; Allen et al., 1993;Capuzza, 2020; Hamilton et al., 2006). Similar studies have beenfar less common in China, as the Chinese picture book market onlydeveloped from the start of the 21 st Century (Xiao, 2021). Onestudy has shown the existence of the traditional gender biases inChinese picture books (Liu & Chen, 2018). Based on the researchgap between these two countries, this study will (1) investigategender representations and narratives in picture books, (2) comparethe similarities and differences between bestselling Britishand Chinese picture books texts from 2010 to 2020. We do so byapplying text mining techniques to analyse gender narratives withinpicture book texts themselves. This follows on from our 2022study where we analysed publisher’s descriptions of texts, ratherthan full text mining of the book’s content (Li et al., 2022).

To detect conflicts in literary texts, this short paper adapts word embedding based sentiment analysis for assigning conflict values to texts. This heuristic approach is able to output relevant verb phrases of a corpus of German novels from the Romantic period and indicate first trends in the corpus.

Graph Literary Exploration Machine (GoLEM) is a new web-based application for literary scholars. The tool allows for named entity relationship analysis, terminology mining, and topic modeling. A strong emphasis is put on the visualisation of results as graphs, time series, maps or scatter plots.

Collective sentiment plays a crucial role in financial markets. In this paper, we present our workflow for creating an index that captures the sentiment at the Berlin Stock Exchange from 1872 to 1930. Based on a historical newspaper, we train three models that provide daily aspect-level sentiment data.

Researchers from Temple University and Emory University will present the results of an ongoing collaboration using topic modeling and word2vec to explore the eco-poetics and eco-politics of a digitized 20th century speculative fiction corpus

The TEITOK REST API lets you interact with your corpus remotely - to upload documents in various formats - automatically converted to TEI/XML; to render search results; to run an NLP pipeline on the server; or download the content of corpus document, treat that content locally with local NLP tools or manual annotation tools, and then upload the results back to the server, where the new or corrected annotations will be incorporated into the original TEI/XML document without destroying any of the potentially complex annotations already in the document.

This 3-hour workshop will introduce span categorization as a method for the machine-annotation of text for various research tasks in the digital humanities. Participants will gain a conceptual understanding of how span categorization differs from entity recognition and complete practical exercises to train a spaCy span categorizer on the LitBank dataset.

Humanistic NLP articulates the multiple challenges facing interactions between NLP and humanities research and offers solutions to bridge this gap. Not merely an application of NLP in DH, it should be considered as an opportunity to engage more humanist scholars in the development and critical appraisal of NLP language resources.

The short presentation introduces the semanticlayertools software package which supports the study of the evolution of scientific fields using network-related methods. An application for the field of astrophysics is presented together with visualization strategies for the emergence of related research fields and the role of specific institutions in thematic clusters.

This study aims to extract people's names from pre-modern Arabic texts. The proposed dictionary-based model eliminates the influence of foreign words and extracts both first names and surnames from three selected corpora. It resolves two fundamental problems faced while performing Named Entity Recognition in Modern Standard Arabic studies.

We present the development of a richly annotated corpus of historical proceedings of the Carniolan Regional Assembly, which was the highest legislative body of the autonomy of the Carniola Herzogtum of the Habsburg Empire. We focus on the procedure used to convert the OCRed scans into Parla-CLARIN TEI encoded documents.

This paper sets out to highlight and address the challenges in working with large audiovisual archives such as the Radio Télévision Suisse collection. The goal of this research is to augment the metadata through Natural Language Processing methods to facilitate access to the archive for the larger public.

The poster/presentation demonstrates a web-based search tool for the Index of Middle English Prose, built using the SRILM language modelling toolkit, which is capable of handling variation in spelling, syntax and lexicon inherent to a prestandardised vernacular such as Middle English.

We will present challenges for a digital edition of an 18th-century gazetteer with 18000 entries on places in the Americas. Challenges range from OCR via applying TEI (dictionary module), orthographic variance, homonym disambiguation and coverage in "matching resources", to the extensive use of relative terminology for places in the prose.

The use of computational approaches with legacy catalogue descriptions enables new insights into their authors' historical perspectives, and provides the means for their evaluation and transformation into contemporary online catalogue records. I discuss my computational research with the historical catalogues of the British Library collection of pre-1500 published books.

The scarcity of machine-readable corpora for Egyptian hieroglyphic texts hinders the digital study of ancient Egyptian texts. Computer-assisted transliteration of the texts will speed up producing such texts and the first step is to find word-boundaries. We present a method for the automatic segmentation of hieroglyphic texts.

We present Pandore, an online toolbox for pre- and post-processing of textual data. This freely accessible platform allows humanities students and researchers, with little or no prior knowledge of computer programming, to perform a series of important and recurrent tasks crucial for the editing, enrichment and mining of digital texts.

Morphological analysis is an indispensable tool for analyzing Japanese text, but it is difficult for ordinary humanities researchers to set up an environment and perform analysis from the command line. Therefore we have developed WebChamame, an online tool that allows users to perform morphological analysis using multiple period-specific UniDic dictionaries.

This contribution deals with Dynamic Topic Modelling of Austrian historical newspapers (found in the ANNO Corpus), focusing on the time period from the 18th to the mid-20th Century. The aim of the investigation is to uncover the development of the topic surrounding migration from and via Southeast Europe to Austria.

Replication studies are generally considered an important way of checking the generalizability of research results. However, in the humanities they are still comparatively rare. We argue that data-driven studies in particular are in need of replication and demonstrate the need in a case study on German academic language.

Our work considers how new advances in pretrained text-to-text generation models might make named entity recognition more accurate, flexible, and streamlined for the digital humanist. We provide an example of how text-to-text generative models can identify mentions of characters, authors, and book names within Goodreads book reviews.

I will present the project of building a database of historical Polish lexicons. The most important dictionaries with Middle Polish material will be digitized in steps: HTR recognition, XML tagging, morphosyntactic tagging. We intend to prepare the API that could be connected with sources of Middle Polish or Latin used in Poland.

We present an overview of the final results of the annotation task of the Emotions in Drama (EmoDrama) research project which are corpora of 18 historical German plays annotated with over 21,000 emotion annotations. We present distribution results, agreement statistics and discuss application areas of this unique corpora.

At present there is no way to search the corpus of Greek papyri for lemmas, or to search for specific grammatical forms of a word. Much less is there a way to search for examples of a grammatical category. Polyphemus comes to solve these shortcomings, and some more. For this purpose we have processed all the papyrus texts from PapyInfo (). This processing is done at the same time as the processing that results in the Callimachus database, which we also present at this DH Congress Congress. I summarize below the procedure by which we obtain our database Polyphemus. A) First we analyze each line of papyrus and differentiate the actual full words from the gaps or non-textual elements. B) Then we identify the complete words and separate them from the fragments. This can sometimes be done because of the editorial criteria used in the original edition, before digitizing. Other times it is necessary to check if the text meets some of the external qualities that define a word in ancient Greek (presence of accentuation, etc.). C) We then proceed to lemmatize each of the words, and determine to which part of speech it corresponds, and what is its morphological analysis. All this is done with the help of the Madrid list, which I will discuss below. For text fragments (incomplete words), we try to see if they can be ascribed to a root. We also separate proper nouns from common nouns. D) Lemma assignment and POS-tagging is performed in two phases. In a first pass we tag the forms with the highest frequency of occurrence. We calculate this from the frequency with which lexical forms were tagged in several manually annotated treebanks (over 700,000 words). We then go on to label all the remaining forms using the Madrid Wordlist. The Madrid Wordlist incorporates information about the dialect in which a form appears, so in case of multiple possible analyses we prefer those belonging to Koine (the Greek form of Papyri Greek) or Attic Greek. Naturally this procedure has the consequence that we reduce the number of multiple analyses for the same form (thus drastically reducing the number of false positives) in exchange for losing the correct POS- tagging for low-frequency forms that coincide with high-frequency ones. E) All this information is transferred to a SQL database, and put in relation with the data on the papyri that we have obtained when creating the Callimachus database. In this way, for each lexical form we obtain a lemma, a non-disambiguated morphological analysis, and a translation or gloss. Each of these parameters can be searched in combination with the more than fifty categories available to us thanks to Callimachus, such as date, origin, category, extension, subject, etc. To date, we have been able to analyze 97% of the complete words, including proper names, which are very numerous. 4. The Madrid Ancient Greek Word List The lemmatization and analysis in Parts Of Speech (POS tagging) is performed by comparing each record in our database with the records of a word list that we have created over the last 3 years, which we have called the Madrid Ancient Greek Wordlist. Most of the Ancient Greek wordlists are evolutions, simplifications, or improvements from the Morpheus list developed by Gregory Crane between 1984 and 1990 (Crane 1991; Celano et al. 2016). Our list also starts with Morpheus, but has been enriched with our own treebank (Aristarchus Treebank, 200,000 words; cf. Riaño 2006), and almost 100,000 proper names from The Lexicon of Greek Personal Names and the Trismegistos repository of papyrological and epigraphic resources. All these data were processed to obtain morphological information. I have manually entered several hundred (mostly irregular) pronominal forms in this list. To complete this list I have processed the digital version of the Greek-English Lexicon of Liddell-Scott-Jones, and extracted all the nominal lemmas; then I have determined the declension of each one of them, and I have proceeded to decline each lemma in its Attic and Ionic form by means of a program we have developed. Then we search for each of these forms in the papyri. The program thus produces over 600,000 lexical forms (many of them already in the Morpheus list). The lemmas are then assigned a translation, or rather a gloss

We use data science methods to examine Colonial Spanish records to analyze the extent of the presence of Afro-descendants in early California. This has been difficult to examine because most Californians of African descent elided their origins in sacramental church records referring to themselves as "gente de razón" or people of reason.

This presentation reports on a case study investigating how NaturalLanguage Processing, a field that applies computational methodssuch as Machine Learning to human-written texts, can supportthe measurement and evaluation of gender biased languagein archival catalogs. Working with English descriptions from thecatalog metadata of the University of Edinburgh’s Archives, wecreated an annotated dataset and classification models that identifygender biases in the descriptions. Conducted with archival data,the case study holds relevance across Galleries, Libraries, Archives,and Museums (GLAM), particularly for institutions withcatalog descriptions in English. In addition to bringing NaturalLanguage Processing (NLP) methods to Archives, we identifiedopportunities to bring Archival Science methods, such as CulturalHumility (Tai, 2021) and Feminist Standpoint Appraisal (Caswell,2022), to NLP. Through this two-way disciplinary exchange,we demonstrate how Humanistic approaches to bias and uncertaintycan upend legacies of gender-based oppression that mostcomputational approaches to date uphold when working with dataat scale.

Humanities textual data is full of references to persons and locations given in various languages. Researchers want to perform queries to retrieve data, in which a certain place or a person is mentioned, irrespective of the language of the text. In this paper, I present how we automatically extract named entities (geolocation information and person references) from textual data and homogenize and store them as Linked Open Data (LOD) with unique identifiers such as the GeoName ID and the GND (Gemeinsame Normdatei) number. Then the plain references in the text are substituted with standoff links to the corresponding RDF resources and the textual document is stored in RDF format. This enables humanities scholars to perform advanced SPARQL queries to collect textual resources containing specific references regardless of the language of the text. Furthermore, the relations between these named entities can be parsed from the text based on ontology definition, dependency graph of sentences, and POS tags to be added to the knowledge graph. Since the citability of the information is crucial for humanities research, this workflow adds the metadata regarding the source document of extracted information to the edges of the knowledge graph using RDF-star. This allows queries for documents containing a certain relationship between entities through SPARQL-star.

This poster presents a suite of Python utilities for interacting with the DICES database of metadata on Greek and Latin epic speeches and for connecting DICES data with a broader ecosystem of linked open data for Classics.

We test the Aesthetic-Affective Potential (AAP), a word-embedding-based predictor for the perceived beauty of words. In a SentiArt context, AAP has also been shown to predict readers' liking of narrative passages. We investigate AAP and valence in the context of a large Dutch corpus of novels and their reviews.

The proposed poster gives an overview on the Gloss-ViBe project which analyses the early medieval Celtic and Latin glosses found in the fragmentory manuscript Vienna, Österreichische Nationalbibliothek, Codex 15298 (olim Suppl. 2698) from different angles including digital humanities, philology and linguistics.

The poster introduces a new, challenging dataset in Hebrew and it seeks to extend automatic metaphor detection capabilities in pre-modern Hebrew. In the poster, we present the major literary characteristics of the corpus, the computational approaches and methods we use and our results.

We present the design and implementation of an NLP pipeline for historical Austrian German (Nestroy's comedies, 1827-1862) including transcriptions of dialectal data. Lacking suitable training data, we created a meta-model that compares the predictions of existing models trained on Modern Standard Germany-German corpora, and calculates a compromise between them.

This contribution presents the difficulties encountered and methods to overcome them when using ready-to-use tools for the elaboration of a processing chain going from OCR to NER and then to the cartographic representation of spaces mentioned in literary texts.

I create and implement a digital workflow for the retrieval and annotation of text relevant for the study of strikes from three major collections of digitized German historical newspapers (late 19th Century).

Can we use NLP to extract information about long-dead languages from secondary sources more than a millennium old? This presentation explores the curation of a dataset based on the Jingdian Shiwen (c. 583), a monumental commentary including over 50,000 phonological annotations on ancient Chinese texts.

This paper combines text reuse detection (using software developed for protein string detection) with genre classification based on the Bert transformer model, to understand changing patterns of genre and intextuality in a large dataset of eighteenth century British and Irish texts.

Digital humanists studying reception history don't yet have infrastructure--technical or human--to match the extensive databases and collaborative networks that support computational analysis of literary texts. We address that gap with new corpora and examples of how computational methods can open up multiple perspectives on literary history.

Our strategy to support filtering the descriptive texts and the transcriptions in Monasterium.net seeks to assign semantic categories for the legal acts they record, such as confiscation, donation, or property sale.

CreoPhon is a pilot database that, for now, only includes Portuguesebased creoles (CreoPhonPt). Its mission is to collect sound phonological data about these languages systematically, to put together a findable, accessible, interoperable, and reusable dataset, and to, by using these data, produce studies that describe so-called "creole phonology" and spread it among the scientific community and the general public.

The increased use of digital methods and tools has been steadily gaining momentum in all areas of life and work since the beginning of the COVID-19 pandemic. Extended methods of data processing and analysis could also be developed and applied in linguistics. In this paper we report on lexical and morpho-pragmatic analyzes of linguistic aspects related to the COVID-19 pandemic in two German corpora.

This presentation will explore the characteristics of humanistic data interfaces and propose a design approach which centres the interrogation of computational modelling procedures. Discussion will revolve around the implementation and analysis of a proof of concept interface, and additionally touch on its relationship to digital humanities computing pedagogy.

We create a Knowledge Graph for Humanities research. Starting with a multidisciplinary dataset of 25,000 OCRed JSTOR papers, we use Deep Learning methods to filter out OCR noise, extract and interrelate research activities, methods and goals, associate them with metadata and transform each paper into approximately 200 RDF triples.

This poster proposal presents a project whose results are the linguistic annotation of ancient Greek bibliographic references with a focus on Named Entity Recognition related to author names and work titles, in order to produce new dynamic text-based tools that are not available in existing indices and catalogs.

This paper presents systems to extract measures for legislators' engagement and absence during committee meetings and the stance and affiliation of non-lawmakers making public comments. We propose a system to track the affiliation of organizations in public comments and whether the organizational representative supports or opposes the bill.

This workshop will discuss the field of stylometry and authorship attribution, and its application to the development of legal evidence. Workshop participants will be invited to participate in a discussion of the development of consensus standards to help improve reliability and to permit evaluation of proposed stylometric evidence.

Uyghur Twelve Muqam is an Intangible Heritage of Humanity. The project "Let data sing the Uyghur Twelve Muqam" aims at using text mining methods to study Muqam lyrics. This paper demonstrates the preliminary results of the project, including topic modeling, sentiment analysis, and social network analysis on Muqam lyrics.

This paper presents our work on training two automatic NER models for ancient Greek using transformer-based models. The models classify the entities into three categories, namely, Person, Location, and Miscellaneous and achieved promising results on test and evaluation datasets.

A previous paper published in this journal proposed a model for evaluating the location of fingermarks on two-dimensional items (de Ronde, van Aken, de Puit and de Poot (2019)). In this paper, we apply the proposed model to a dataset consisting of letters to test whether the activity of writing a letter can be distinguished from the alternative activity of reading a letter based on the location of the fingermarks on the letters. An experiment was conducted in which participants were asked to read a letter and write a letter as separate activities on A4- and A5-sized papers. The fingermarks on the letters were visualized, and the resulting images were transformed into grid representations. A binary classification model was used to classify the letters into the activities of reading and writing based on the location of the fingermarks in the grid representations. Furthermore, the limitations of the model were studied by testing the influence of the length of the letter, the right- or left-handedness of the donor and the size of the paper with an additional activity of folding the paper. The results show that the model can predict the activities of reading or writing a letter based on the fingermark locations on A4-sized letters of right-handed donors with 98 % accuracy. Additionally, the length of the written letter and the handedness of the donor did not influence the performance of the classification model. Changing the size of the letters and adding an activity of folding the paper after writing on it decreased the model's accuracy. Expanding the training set with part of this new set had a positive influence on the model's accuracy. The results demonstrate that the model proposed by de Ronde, van Aken, de Puit and de Poot (2019) can indeed be applied to other two-dimensional items on which the disputed activities would be expected to lead to different fingermark locations. Moreover, we show that the location of fingermarks on letters provides valuable information about the activity that is carried out.

The scientific, clinical, and pedagogical significance of devising methodologies to train nonprofessional subjects to recognize diagnostic visual patterns in medical images has been broadly recognized. However, systematic approaches to doing so remain poorly established. Using mammography as an exemplar case, we use a series of experiments to demonstrate that deep learning (DL) techniques can, in principle, be used to train naïve subjects to reliably detect certain diagnostic visual patterns of cancer in medical images. In the main experiment, subjects were required to learn to detect statistical visual patterns diagnostic of cancer in mammograms using only the mammograms and feedback provided following the subjects’ response. We found not only that the subjects learned to perform the task at statistically significant levels, but also that their eye movements related to image scrutiny changed in a learning-dependent fashion. Two additional, smaller exploratory experiments suggested that allowing subjects to re-examine the mammogram in light of various items of diagnostic information may help further improve DL of the diagnostic patterns. Finally, a fourth small, exploratory experiment suggested that the image information learned was similar across subjects. Together, these results prove the principle that DL methodologies can be used to train nonprofessional subjects to reliably perform those aspects of medical image perception tasks that depend on visual pattern recognition expertise.