NLP for the Ancient World
Showing 5 of 7 projects. Click any project card for scope, mentors, and proposal studio.
Mentors: Student: Eleftheria
<p>NLP is severely lacking in meaningful functionalities for Germanic languages. Normalization, POS tagging and stemming modules (all significant parts of modern NLP) seem to be absent from the core, rendering the use of the tool as part of significant research projects non-viable. The proposed progress aims to extend the current functionalities of the Germanic family.</p>
Mentors: Student: James Gawley
<p>The CLTK features the most sophisticated algorithm available for lemmatizing classical Latin. Lemmatization is the process by which inflected word-forms are grouped together under their dictionary headings. This allows us to gather accurate word-usage statistics, analyze authorship, and model subject matter in classical corpora. However the CLTK lemmatizer is not currently able to identify synonyms for a given word, or suggest translations into other languages.</p> <p>I propose to modify the existing CLTK lemmatizer to look up synonyms and translations for Latin and Greek. I will adapt CLTK’s unique ‘backoff’ approach to lemmatization in order to measure the probability of each possible synonym and/or translation for a target word given its context. Further, I propose to incorporate vector models for Latin and Greek based on word embeddings trained using the word2vec algorithm. Once synonyms, translations, and vector models are incorporated into CLTK, users will be able to perform cutting-edge tasks like sentence length document alignment. This will open new horizons for digitally assisted classical scholarship.</p>
Mentors: Student: Andrew Deloucas
<p>This project focuses on integrating Cuneiform Digital Library Initiative (CDLI) corpora into the Classical Language Toolkit (CLTK). Currently, CLTK houses several functions developed by Dr. Willis Monroe; the difficulty in utilizing these functions is due to the variables having to be presented in a reconstructed normalized form of Akkadian, which is not how tablets are either written traditionally or stored by CDLI. The goal of this project is to enable CLTK to reconstruct a normalized spelling out of CDLI’s ATF text and thus create data available for two fundamental uses: 1) allowance of individuals to learn and practice Akkadian with real and novel reading exercises; and 2) with further class development, be analyzed on a mass scale.</p>
Mentors: Student: Natasha Voake
<p>Old and Middle French are hardly studied outside of a limited network of French universities. Implementing NLP functionality to these languages would make it easier to study them and access the rich literature and culture expressed in them, famous examples of which include the "Chanson de Roland", Chrétien de Troyes' Arthurian legends, Marie de France's "Lais", Christine de Pizan's writings, etc.</p> <p>This project aims to extend basic CLTK functionality to Old and Middle French texts between c.900 and c.1500 CE, by implementing a tokenizer, stopwords, named entity recognition, a PoS tagger, and a lemmatizer with English translations for as many words as possible. Data from which the above will be sourced will be from texts licensed under creative commons licenses, which have been transcribed and digitized. For example, a number of Old French texts from the BNF's 19th century editions have been digitized and made available at gallica.fr. Lemmas will be sourced from Godefroy's 1901 "Lexique de l'Ancien Français" and the "Dictionnaire Electronique de Chrétien de Troyes", which has the advantage of English-language definitions.</p>
Mentors: Student: Charles Pletcher
<p>I plan to build a flexible platform for teachers to annotate texts in the CLTK Archive. The purpose of the annotations is twofold: first, to guide the students in their (re)readings of the text; and second, to provide a ready-to-hand reference to which students and teachers can return.</p> <p>Teachers will be able to enter annotations either by uploading a list of references and notes or by entering their annotations directly through a WYSIWYG interface built on top of <a href="https://draftjs.org/" target="_blank">Draft.js</a>. Annotations will be searchable via <a href="https://www.elastic.co/products/elasticsearch" target="_blank">Elasticsearch</a>, taggable, and, if time allows, versioned.</p>