Natural Language Processing (NLP) – University of Copenhagen

Natural Language Processing (NLP)

NLP is a subfield of artificial intelligence, concerned with automatically processing as well as understanding text, typically using statistical models and machine learning. The resulting models can be used for applications such as correcting grammatical errors, summarising documents, answering questions or translating texts.

In our research group, we are interested in traditional core as well as emerging topics in natural language processing. We focus on researching models that require little direct input from humans to work well, and models that work well across languages. We are further interested in exploiting different modalities, for instance visual and textual data referring to the same event from captioned images, or speech and image sequences from videos.

Our work has been reported on in several international newspapers, nominated for and awarded with several best paper awards at scientific conferences, and established Copenhagen as an international center for NLP. Since its inception in 2013, the group grown to three faculty members, and is now ranked in the top 10 of NLP groups worldwide and 2nd in Europe, according to CSRankings.

Research

Anders Søgaard conducts research in core topics in natural language processing and applied machine learning, including (cross-lingual and cross-domain) transfer- and multi-task learning, adversarial learning, and reinforcement learning.

Isabelle Augenstein’s main research interests are natural language understanding and learning with limited labelled data. This includes emerging topics such as stance detection and fact checking; as well as representation, few-shot and weakly supervised learning.

Desmond Elliott’s research focuses on models that solve problems by integrating vision and language. He is interested in multimodal machine translation, image captioning, and multilingual video understanding.

Research Projects

Learning Multilingual Natural Language Understanding from Transfer & Reinforcement (2018-2022)
Google Focused Research Award
Led by Anders Søgaard

Building Emotionally Intelligent Dialogue Systems (2018-2021)
BotXO Research Award
Led by Anders Søgaard

Automated Writing Assistant with Artificial Intelligence (2018-2021)
Industrial PhD Grant
Led by Anders Søgaard

Advanced Machine Learning for Automated Omni-Channel Support (2018-2020)
Grand Solutions grant from Innovation Fund Denmark
Led by Stephen Alstrup
More information (in Danish)

NLPL: Nordic Language Processing Laboratory (2017-19)
Infrastructure grant from NeIC
Led by Bjørn Lindi
More information

Paper-based paper search (2018)
AI2 Faculty Award
Led by Anders Søgaard

Digital Disinformation (2016-18)
Research grant from the Carlsberg Foundation
Led by Rebecca Adler-Nissen
More information

ReProsis: Real Time Big Data Product Analysis – Product Management System in International Markets (2016-18)
Research project funded by Eurostars
Led by Dirk Hovy and Isabelle Augenstein
More information

Interactive text simplification for dyslexics (2015-18)
Research project supported by Trygfonden
More information

Previous Research Projects

From dogma to data (2015-18)
Research project supported by the Danish Research Council
Led by Anders Søgaard with Henrik Palmer Olsen, iCourts, Faculty of Law, University of Copenhagen
More information

Uncertain archives (2015-18)
Research project supported by the Danish Research Council
Led by Kristin Veel
More information

Interactive text simplification for dyslexics (2015-18)
Research project supported by Trygfonden
More information

From a thesaurus to a Danish FrameNet (2016-17)
Infrastructure grant from the Carlsberg Foundation

Semantic processing across domains (2014-17)
Research project supported by the Danish Research Council
Led jointly with Bolette S. Pedersen
More information (in Danish)

LOWLANDS: Parsing low-resource languages and domains (2013-17)
Research project (ERC Starting Grant) supported by the European Research Council
Led by Anders Søgaard
More information

Finding Waldo in a haystack of informal writing styles (2016-17)
Research grant from the Data Transparency Lab
More information

RELIP: Reading between the lines (2016)
Seed money, Patient At Home
More information (in Danish)