I'm a Natural Language Processing (NLP) scientist and engineer, working on NLP and LLM applications in specialised domains such as defence and medicine, with a strong focus on multilinguality. I'm also drawn to the intersection of NLP/AI with the humanities and ethics.
Career
- 2024–present
NLP scientist and engineer, Textgain Multilingual NLP for specialised domains: harmful-content detection (Calico) and bias-resilient intelligence analysis (SIMBA).
- 2020–2024
Research fellow, ARC Training Centre in Cognitive Computing for Medical Technologies, University of Melbourne Automated quality assessment of medical evidence and COVID-SEE, with Karin Verspoor and Tim Baldwin.
- 2016–2020
Postdoc, CLiPS, University of Antwerp Clinical NLP, machine reading comprehension and interpretability, with Walter Daelemans.
- 2012–2016
PhD, Computational Linguistics, University of Groningen Thesis on word representations, advised by Gertjan van Noord and Ivan Titov.
- 2010–2012
MSc, Language and Communication Technologies, U. of Groningen and U. de Lorraine
- 2003–2008
University degree, Translation studies, University of Ljubljana
Selected projects
- Calico LLMsDevelopment of multilingual LLMs for harmful-content detection.
- SIMBAMulti-agent AI that helps defence analysts spot bias and logical fallacies in intelligence sources.
- Evidence gradingAutomatic quality assessment of medical evidence for systematic reviews.
- COVID-SEEA scientific evidence explorer for COVID-19 research.
Selected publications
- Detecting Spin in Clinical Trials with Large Language Models
arXiv preprint arXiv:2610.11845, 2026. - Zero- and Few-Shot Prompting of Generative Large Language Models Provides Weak Assessment of Risk of Bias in Clinical Trials.
Research Synthesis Methods, 2024. - Automating Quality Assessment of Medical Evidence in Systematic Reviews: Model Development and Validation Study
Journal of Medical Internet Research, 25: e35568, 2023. - Promoting Fairness in Classification of Quality of Medical Evidence
BioNLP, ACL, 2023. - Mapping probability word problems to executable representations.
EMNLP (long paper), 2021. - CliCR: A Dataset of Clinical Case Reports for Machine Reading Comprehension.
NAACL (long paper), 2018. bibtexcode
Show all 37 publications
2026
- Detecting Spin in Clinical Trials with Large Language Models
arXiv preprint arXiv:2610.11845, 2026.
2024
- Zero- and Few-Shot Prompting of Generative Large Language Models Provides Weak Assessment of Risk of Bias in Clinical Trials.
Research Synthesis Methods, 2024.
2023
- Uncertainty Estimation for Debiased Models: Does Fairness Hurt Reliability?
IJCNLP-AACL, 2023. - Promoting Fairness in Classification of Quality of Medical Evidence
BioNLP, ACL, 2023. - Analysis of predictive performance and reliability of classifiers for quality assessment of medical evidence revealed important variation by medical area
Journal of Clinical Epidemiology, 2023. - Automating Quality Assessment of Medical Evidence in Systematic Reviews: Model Development and Validation Study
Journal of Medical Internet Research, 25: e35568, 2023.
2022
- Predicting Publication of Clinical Trials Using Structured and Unstructured Data: Model Development and Validation Study
Journal of Medical Internet Research, 24 (12): e38859, 2022.
2021
- Mapping probability word problems to executable representations.
EMNLP (long paper), 2021. - Impact of detecting clinical trial elements in exploration of COVID-19 literature
HealthNLP Workshop, 2021. - Are we there yet? Exploring clinical domain knowledge of BERT models
BioNLP, NAACL, 2021. - Contextual explanation rules for neural clinical classifiers
BioNLP, NAACL, 2021. - Scalable Few-Shot Learning of Robust Biomedical Name Representations
BioNLP, NAACL, 2021. - Integrating Higher-Level Semantics into Robust Biomedical Name Representations
Workshop on Health Text Mining and Information Analysis (LOUHI), EACL, 2021. - Conceptual Grounding Constraints for Truly Robust Biomedical Name Representations
EACL, 2021. - Brief description of COVID-SEE: The Scientific Evidence Explorer for COVID-19 Related Research
ECIR, 2021.
2020
- Improved Topic Representations of Medical Documents to Assist COVID-19 Literature Exploration
NLP COVID-19 Workshop at EMNLP, 2020. - COVID-SEE: Scientific Evidence Explorer for COVID-19 Related Research
arXiv preprint arXiv:2008.07880, 2020. - Distilling neural networks into skipgram-level decision lists
arXiv preprint arXiv:2005.07111, 2020.
2019
- Why can't memory networks read effectively?
arXiv preprint arXiv:1910.07350, 2019. code - Unsupervised Concept Extraction from Clinical Text through Semantic Composition.
Journal of Biomedical Informatics, 2019. code
2018
- Revisiting neural relation classification in clinical notes with external information.
Workshop on Health Text Mining and Information Analysis (LOUHI), EMNLP, 2018. bibtexpostercode - Rule induction for global explanation of trained models.
Analyzing and interpreting neural networks for NLP (BlackBoxNLP), workshop at EMNLP, 2018. bibtexpostercode - Patient representation learning and interpretable evaluation using clinical notes.
Journal of Biomedical Informatics, 2018. bibtex - CliCR: A Dataset of Clinical Case Reports for Machine Reading Comprehension.
NAACL (long paper), 2018. bibtexcode
2017
- Unsupervised patient representations from clinical notes with interpretable classification decisions.
NIPS Machine Learning for Health Workshop, 2017. bibtexposter - Unsupervised Context-Sensitive Spelling Correction of English and Dutch Clinical Free-Text with Word and Character N-Gram Embeddings.
CLIN Journal, 2017. bibtexcode - Unsupervised Context-Sensitive Spelling Correction of Clinical Free-Text with Word and Character N-Gram Embedding.
BioNLP, 2017. bibtexcode - A Short Review of Ethical Challenges in Clinical Natural Language Processing.
First Workshop on Ethics in NLP, EACL, 2017. posterbibtex
2016
- Using Distributed Representations to Disambiguate Biomedical and Clinical Concepts.
BioNLP, 2016. bibtexcode - Empirical studies on word representations.
PhD thesis, 2016. bibtex - Bilingual Learning of Multi-sense Embeddings with Discrete Autoencoders.
NAACL (long paper), 2016. bibtexslidescode
2015
- Word Representations, Tree Models and Syntactic Functions.
arXiv preprint arXiv:1508.07709, 2015. bibtexcode - GLAD: Groningen Lightweight Authorship Detection.
Uncovering Plagiarism, Authorship and Social Software Misuse, CLEF, Author Identification challenge, 2015. code - An investigation into language complexity of World-of-Warcraft game-external texts.
arXiv preprint arXiv:1502.02655, 2015. bibtex
2014
2013
- Semantic Mapping for Lexical Sparseness Reduction in Parsing.
ESSLLI Extrinsic Parse Improvement Workshop, 2013. bibtex
<2013
Supervision
- Current: two students from the Master of AI at KU Leuven (industry theses on uncertainty in LLMs and on multi-agent architectures).
- Co-advised to completion:
- Siyang Wang and Fan Ye, MSc, medical NLP, University of Melbourne.
- Madhumita Sushil, PhD, University of Antwerp: interpretability and document representation learning.
- Pieter Fivez, PhD, University of Antwerp: lexical normalisation and modeling of lexical variability.
- Tjaš Ajdovec, MSc, University of Ljubljana: detecting misleading reporting in scientific literature using LLMs.
Selected talks
- Automating risk-of-bias assessment with generative AI Long oral talk, Global Evidence Summit, Prague, 2024.
- Maske ne obvarujejo pred virusi. Ali pač? Avtomatsko ugotavljanje kakovosti izsledkov v medicini Vabljeno predavanje, seminar JOTA, Fakulteta za računalništvo in informatiko Univerze v Ljubljani, 2023. povzetekvideo
- Automated quality assessment of medical evidence to support systematic reviewing. NLP reading group talk, Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, 2023.
- Automated quality assessment of medical evidence to support systematic reviewing. Invited talk, CLiPS Colloquium, University of Antwerp, 2022.
- Automated quality assessment of medical evidence to support systematic reviewing. Invited talk, UMass BioNLP, 2022.
- Fill the gap: Machine reading comprehension for medicine. Invited talk at NLP Meetup Belgium, May 2019.
- Spelling correction with word and character n-gram embeddings. Invited talk at Vectors and Linguistics: a workshop on word embeddings, University of Leiden, March 2018.
- Representation learning for words. Guest lecture at the Current trends in AI master course, Free University of Brussels, 2017.
Show all 51 talks
2024
- Automating risk-of-bias assessment with generative AI Long oral talk, Global Evidence Summit, Prague, 2024.
2023
- Maske ne obvarujejo pred virusi. Ali pač? Avtomatsko ugotavljanje kakovosti izsledkov v medicini Vabljeno predavanje, seminar JOTA, Fakulteta za računalništvo in informatiko Univerze v Ljubljani, 2023. povzetekvideo
- Automated quality assessment of medical evidence to support systematic reviewing. NLP reading group talk, Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, 2023.
- Using Machine Learning and Natural Language Processing to Structure Medical Evidence and Grade its Quality. Mobilizing Computable Biomedical Knowledge (MCBK) meeting, 2023.
2022
- Automated quality assessment of medical evidence to support systematic reviewing. Invited talk, CLiPS Colloquium, University of Antwerp, 2022.
- When to trust a classifier for quality assessment of medical evidence? International Collaboration for the Automation of Systematic Reviews (ICASR), Cologne, Germany, 2022.
- Automated quality assessment of medical evidence to support systematic reviewing. Invited talk, UMass BioNLP, 2022.
- Robustness and practical applicability of EvidenceGRADEr. Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, short talk, 2022.
2021
- Automated quality assessment of medical evidence. Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, short talk, 2021.
- Automated quality assessment of medical evidence. Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, long talk, 2021.
- Automated quality assessment of medical evidence. Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, short talk, 2021.
2020
- Stream 4: Synthesising, appraising and exploring medical evidence. ITTC Healthcare Symposium, 2020.
- Creating a dataset for automated quality assessment of medical evidence (part 2). Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, short talk, 2020.
- I want to know what attention is (I want you to show me). ARC Training Centre in Cognitive Computing for Medical Technologies, machine learning reading group., 2020.
- Creating a dataset for automated quality assessment of medical evidence (part 1). Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, short talk, 2020.
- COVID-SEE: Scientific Evidence Explorer for COVID-19. Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, long talk, 2020.
- Real-time clinical decision support. Seminar of ARC Training Centre in Cognitive Computing for Medical Technologies, short talk, 2020.
2019
- Machine reading for medicine. Closing workshop of the Accumulate project, KULeuven, 2019.
- Why memory networks can’t read effectively? CLiPS, University of Antwerp, 2019.
- Sequence models for probability word problems. CLiPS, University of Antwerp, 2019.
- Fill the gap: Machine reading comprehension for medicine. Invited talk at NLP Meetup Belgium, May 2019.
- Machine reading comprehension and word problem solving (with external knowledge). Annual meeting of CLiPS, 2019.
- Memories are made of this: A primer on memory networks for QA. CLiPS, University of Antwerp, 2019.
- Revisiting neural relation classification in clinical notes with external information. CLIN, 2019.
2018
- Technology developed at CLiPS. Accumulate industrial meeting, September 2018.
- Spelling correction with word and character n-gram embeddings. Invited talk at Vectors and Linguistics: a workshop on word embeddings, University of Leiden, March 2018.
- Technology developed at CLiPS. Accumulate industrial meeting, March 2018.
2017
- Clinical Machine Comprehension with Case Reports. ATILA, 2017.
- Clinical Machine Comprehension Using Case Reports. (poster) American Medical Informatics Association (AMIA) Annual Symposium, 2017.
- What is attention in NNs? (with two examples) A 20-min tutorial, CLiPS, University of Antwerp, 2017.
- Representation learning for words. Guest lecture at the Current trends in AI master course, Free University of Brussels, 2017.
- Clinical Case Reports Dataset for Machine Reading. (poster) 27th meeting of Computational Linguistics in the Netherlands (CLIN), 2017.
2016
- The challenges in concept detection for clinical texts. Accumulate industrial meeting, 2016.
- Towards clinical language understanding. ATILA, 2016.
- Clinical language processing: the first steps Annual meeting of CLiPS, 2016.
2015
- Inducing multi-sense word representations multilingually. ATILA and CLIN, 2015.
- Who's the bad guy? OlympIKade: Informatiekunde Matchingsdag, 2015.
- Presentation of E. Bender's (2011) On Achieving and Evaluating Language-Independence in NLP. RUG Computational Linguistics reading group, 2015.
- Overview of Learning From Data’s Final Project: Author Verification. With Malvina Nissim. RUG Computational Linguistics reading group, 2015.
- Tree models, syntactic functions and word representations. The 25th meeting of Computational Linguistics in the Netherlands (CLIN), 2015.
2014
- From perceptrons to word embeddings (a high-level introduction). RUG Computational Linguistics reading group, 2014.
- Extending Hidden Markov (tree) models for word representations. (poster) 23rd annual Belgian-Dutch Conference on Machine Learning (BENELEARN), 2014. abstract
- How to write a master's thesis: a computational linguist's view. RUG Research Master's in Linguistics meeting, April 2014, March 2015.
- Dependency-tuned word clusters for Dutch. The 24th meeting of Computational Linguistics in the Netherlands (CLIN), 2014.
2013
- Reading group presentation on Reddy et al. 2011 paper on Dynamic and Static Prototypes For Semantic Composition. RUG Computational Linguistics reading group, 2013.
- Semantic Mapping for Lexical Sparseness Reduction in Parsing. ESSLLI Extrinsic Parse Improvement Workshop, 2013.
- The Brown et al. 1992 Clustering. RUG Computational Linguistics reading group, 2013.
- Semantic Mapping for Lexical Sparseness Reduction in Parsing. New Frontiers in Parsing and Generation Workshop, 2013.
- Lexical Association Analysis For Semantic-Class Feature Enhancement In Parsing. (poster) 23rd meeting of Computational Linguistics in the Netherlands (CLIN), 2013.
<2013
- Resolving PP-attachment ambiguity by distributional semantic modeling in the context of parsing of French. RUG Computational Linguistics reading group, 2012.
- The SSJ corpus in the context of Slovene reference corpora. With Olga Yeroshina Pobirk. 7th International Conference Practical Applications in Language and Computers, 2009.
- →Talks in Slovene
Teaching
Computational Linguistics at the University of Antwerp (2017–2020, with Walter Daelemans); Learning from data (2013–2015) and Corpus linguistics (2011, 2013) at the University of Groningen.
Show course details
2017/2018, 2018/2019, 2019/2020
- Computational Linguistics, UA_2010FLWTAA. With prof. dr. Walter Daelemans
- POS-tagging & Minimum Edit Distance
- Syntactic Analysis & Parsing
- Semantic Role Labeling & Frame Semantics
2014/2015
- Learning from data. With dr. Malvina Nissim
- final project: Authorship verification
- topics from 2013/2014
2013/2014
- Learning from data. With prof. dr. Gertjan van Noord
- Introduction to Weka, The Perceptron
- K-means
- Brown Clustering
- Linear Regression
- final project: Movie revenue prediction from reviews
<2013/2014
- Corpustaalkunde (Corpus linguistics), 2013. Assisting dr. Gosse Bouma
- Corpustaalkunde (Corpus linguistics), 2011. Assisting dr. Gosse Bouma