Language Technology: Research and Development Workshop

The final workshop will take place on Zoom. Students and staff will get information about the Zoom rooms via email. If anyone else is interested in attending this workshop, please contact Sara Stymne (sara.stymne@lingfil.uu.se) for details.

Each student will have 12 minutes for presentation, followed by 3 minutes for questions.

Zoom room 1 Zoom room 2

9.30 Sara Welcome and introduction

Embeddings: gender bias and grammar: Ali

9.45 Ziyang Have Attention Heads in BERT Learned Phrase Grammar?

10.00 Meichun Investigating Gender Bias Hidden Behind Chinese Adjectives in Word Embeddings

10.15 Maria-Elena Reducing gender bias in NMT with debiased word embeddings

10.30 BREAK

Parsing and readability: Sara BERT and sentiment: Daniel

11.00 Bjarki Cross-lingual dependecy parsing methods for Icelandic past and present Xi Generating knowledge graph embedding using KG-BERT on the DrugBank with probablistic negative sampling

11.15 Yifei Impact of Language Distances and Data Sizes on cross-lingual Parsing Pochun Method Exploration: Relation Extraction Using BERT

11.30 Antonia Investigation of Transfer Languages for Parsing Latin: Italic Branch vs. Hellenic Branch Giacomo Steam game reviews sentiment analysis with a CNN Twitter-enriched model

11.45 Gustav A New Readability Measure Based on Genre Classification Xindi A Comparative Study of Neural Network Models for Chinese Movie Reviews: In the Aspect of Data Classification and Segmentation Method

12.00 LUNCH BREAK

WSD, NE, offensive language: Sara

13.00 Huiling Evaluating Contextualized Multilingual Language Models in Word-in-Context Disambiguation

13.15 Xingran Cross-lingual Word Sense Disambiguation using mBERT Embeddings with Syntactic Dependencies

13.30 Ahmed Arabic Named Entity Recognition using contextualized word embeddings

13.45 Sebastian Preprocessing and the Detection of Offensive Language in German Tweets: an Exploration How Different Methods Interact with Different Classifiers

14.00 BREAK

Cross-lingual sentiment and tagging: Daniel

14.30 Sijia What Multilingual BERT Learns from Back-Translation for Sentiment Classification Tasks

14.45 Naomi You Can Take Your Complaint to a Different Language: Sentiment Analysis Using Cross-Lingual Data and Convolutional Neural Networks in a Low Resource Situation

15.00 Melvin Sentimix Dataswitch: Comparing Diferente Spanglish Data Sets on Models

15.15 Harm Whit's the Richt Pair o Speech: PoS tagging for Scots

15.30 Sara Wrapping up

	Zoom room 1		Zoom room 2
9.30	Sara	Welcome and introduction
	Embeddings: gender bias and grammar: Ali
9.45	Ziyang	Have Attention Heads in BERT Learned Phrase Grammar?
10.00	Meichun	Investigating Gender Bias Hidden Behind Chinese Adjectives in Word Embeddings
10.15	Maria-Elena	Reducing gender bias in NMT with debiased word embeddings
10.30	BREAK
	Parsing and readability: Sara		BERT and sentiment: Daniel
11.00	Bjarki	Cross-lingual dependecy parsing methods for Icelandic past and present	Xi	Generating knowledge graph embedding using KG-BERT on the DrugBank with probablistic negative sampling
11.15	Yifei	Impact of Language Distances and Data Sizes on cross-lingual Parsing	Pochun	Method Exploration: Relation Extraction Using BERT
11.30	Antonia	Investigation of Transfer Languages for Parsing Latin: Italic Branch vs. Hellenic Branch	Giacomo	Steam game reviews sentiment analysis with a CNN Twitter-enriched model
11.45	Gustav	A New Readability Measure Based on Genre Classification	Xindi	A Comparative Study of Neural Network Models for Chinese Movie Reviews: In the Aspect of Data Classification and Segmentation Method
12.00	LUNCH BREAK
	WSD, NE, offensive language: Sara
13.00	Huiling	Evaluating Contextualized Multilingual Language Models in Word-in-Context Disambiguation
13.15	Xingran	Cross-lingual Word Sense Disambiguation using mBERT Embeddings with Syntactic Dependencies
13.30	Ahmed	Arabic Named Entity Recognition using contextualized word embeddings
13.45	Sebastian	Preprocessing and the Detection of Offensive Language in German Tweets: an Exploration How Different Methods Interact with Different Classifiers
14.00	BREAK
	Cross-lingual sentiment and tagging: Daniel
14.30	Sijia	What Multilingual BERT Learns from Back-Translation for Sentiment Classification Tasks
14.45	Naomi	You Can Take Your Complaint to a Different Language: Sentiment Analysis Using Cross-Lingual Data and Convolutional Neural Networks in a Low Resource Situation
15.00	Melvin	Sentimix Dataswitch: Comparing Diferente Spanglish Data Sets on Models
15.15	Harm	Whit's the Richt Pair o Speech: PoS tagging for Scots
15.30	Sara	Wrapping up