Language Technology: Research and Development Workshop

The final workshop will mainly take place on Campus. There will be two Zoom sessions (3a and 4a), which will be broadcast in a room, as well as available directly on Zoom. The Zoom link is avaiable in Studium. (If anyone else is interested in attending the Zoom session, please contact Sara Stymne ( for the link.)

We welcome guests who are not attending the course. However, due to restrictions, we might need to limit access to the class rooms. In that case, priority is of course given to students taking the course.

Note that this schedule is preliminary, and it is subject to change. A few more talks might have to be given on Zoom, which will then affect also other talks, to keep the Zoom talks to specific sessions.

Each student will have 12 minutes for presentation, followed by 3 minutes for questions.
Geijersalen 6-1023 7-0043
9.00SaraWelcome and introduction
Plenary session 1: Sara
9.05K├ĄtriinFinding the best language model - approach combination: cross-lingual methods for intent detection in Estonian
9.20LingqingEvaluating Recurrent Neural Network Semantics: Presupposed Taxonomies
9.35MartinaAuthorship Attribution for Swedish Novels
9.50WenwenHow Spelling Normalization Affects Historical Text Classification
10.05OreenMultilingual Named Entity Recognition via Data Augmentation
10.20BREAK with coffee in Matikum
Session 2a, Zoom talks: PaolaSession 2b: Eva
10.50ChuchuInvestigating the Chinese-English Shared Translation Effect on GPT2ClairePOS tagging for historical Swedish using KB-BERT
11.05PalomaIs Chile a country or a pepper? Spanish Pre-Trained BERT semantic understanding of polysemous wordsEmmaLexical Semantic Change Detection for Historical Swedish Using Topic Modelling
11.20YongchaoA systematic comparison between BERT and ConceptNet Numberbatch on word association propertiesJae EunSwedish Grammar Error Correction Using Neural Networks And Data Supplement Methods
11.35EvaThe Age of Context-Aware Language Embeddings: BERTje and mBERTFlorPatronizing and Condescending Language (PCL) Detection Across Machine Learning Models Using Frequency Based and Word Embedding Representations
Session 3a, Zoom talks: EvaSession 3b: Sara
13.00ChenImplementation and evaluation cross three methods in Automated Cognate JudgementZheComparative analysis of effectiveness of transfer learning and meta learning in automatic Chinese sarcasm detection
13.15ZimingTemporal Classification Based on Supervised LearningAngelikiTowards a POS-Tagger for Griko using Transfer Learning from Greek and Italian
13.30Kai YanImproving Extremely Low-Resource Neural Machine Translation via Unsupervised Bitext Mining: A Case Study of CantoneseJamesThere Is Something About Czech
Session 4a, Campus+Zoom: EvaSession 4b, Zoom talks: Sara
13.55Nikolina (Campus)Machine Learning techniques and Stylometric features on Medieval Greek Text ClassificationYifanMultilingual Idiomaticity Detection: Transfer Learning with English and German for Portuguese Idiom Detection Task with mBERT
14.10Klaudia (Zoom)Evaluating Neural Network Semantics: Detecting Taxonomy with EmbeddingsSiyiUsing Entity Annotation Projection for Multilingual Complex Named Entity Recognition
14.25Viktorija (Zoom)The Impact of Morphological Richness on Word Association Norms in Word EmbeddingsKrisMaltese Named Entity Recognition with Cross-Lingual Conditional Random Fields
14.40BREAK with coffee in Matikum
Plenary session 5, Zoom talks: Paola
15.10JiayiSyntactic Behaviour of Chinese Neural Network Language Models
15.25LauraOCR Quality Evaluation for Natural Language Processing Tasks on Historical Swedish Newspapers
15.40RafalUsing Unaligned Monolingual Data to Build Multilingual and Improve Monolingual Reverse Dictionaries
15.55SaraWrapping up
16.00Social event - CANCELLED