---
title: "Luxembourg — research institutions, scientists and papers"
description: "Luxembourg: 2 institutions with papers in our library in 1 agglomerations, 8 people with a page, 7 works, 285 citations."
canonical: "https://research.elamur.ai/free/en/countries/luxembourg/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/countries/luxembourg/)

[home](https://md.research.elamur.ai/free/en/) / [countries](https://md.research.elamur.ai/free/en/countries/) / Luxembourg

# Luxembourg

Country 2 institutions in 1 agglomeration · [all countries](https://md.research.elamur.ai/free/en/countries/)

[**2** institutions with a page](https://md.research.elamur.ai/free/en/countries/luxembourg/#institutions) · [**1** agglomeration](https://md.research.elamur.ai/free/en/countries/luxembourg/#agglomerations) · [**8** people with a page](https://md.research.elamur.ai/free/en/countries/luxembourg/#authors) · [**7** works in our library](https://md.research.elamur.ai/free/en/countries/luxembourg/#works) · [**285** citations of their works](https://md.research.elamur.ai/free/en/countries/luxembourg/#works)

## About

### Profile

Luxembourg appears on this site through a small, tightly focused body of research in language technology. Its work here comes from two institutions in a single agglomeration, [Luxembourg](https://md.research.elamur.ai/free/en/agglomerations/luxembourg/). The [University of Luxembourg](https://md.research.elamur.ai/free/en/universities/university-of-luxembourg/), based in the city of Luxembourg, accounts for nearly all of the country's works and citations. The [Luxembourg Institute of Science and Technology](https://md.research.elamur.ai/free/en/labs/luxembourg-institute-of-science-and-technology/), a research institute in Belvaux, contributes to part of that output. Both belong to the same urban region, so Luxembourg's research here is in effect the research of one hub.

The subject matter is narrow and coherent. Nearly every work deals with how computers process, translate, generate or evaluate human language. The dominant concern is languages that existing systems serve poorly: Luxembourgish, Mooré, and the many spoken varieties of Arabic. Alongside these sits one very widely cited benchmark for measuring the abilities of large language models. The country's place among [all countries](https://md.research.elamur.ai/free/en/countries/) on this site reflects that profile: a modest number of works, with one paper drawing most of the attention.

### History

Among the works here, Luxembourg's research begins in 2025. That year produced the two most cited papers. The first is a large collaborative benchmark for testing advanced AI systems. The second is a shared task on Arabic text generation and the detection of machine-written Arabic. Both are community efforts: many researchers contribute to a common test, rather than a single group reporting its own method.

In 2026 the work broadens and becomes more specific to individual languages. That year brings studies of machine translation for Luxembourgish and other low-resource languages, speech-to-speech translation between French and Mooré, a second Arabic shared task centred on dialects, and a study of how word meanings change over time. The 2026 papers are also the ones where the Luxembourg Institute of Science and Technology joins the university as an affiliation. The line of work moves from broad evaluation and shared tasks in 2025 toward focused studies of particular languages and particular methods in 2026.

### Research profile

The central field is [linguistics and languages](https://md.research.elamur.ai/free/en/shelves/linguistics/), closely joined by [machine translation](https://md.research.elamur.ai/free/en/shelves/machine-translation/) as a field. Smaller shares fall under [natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/) and, through the benchmark work, [mathematics and statistics](https://md.research.elamur.ai/free/en/shelves/mathematics/).

By topic, the strongest strand is [machine translation](https://md.research.elamur.ai/free/en/topics/machine-translation/) for [low-resource languages](https://md.research.elamur.ai/free/en/topics/low-resource-languages/). Three studies anchor it. In [Are Small Language Models the Silver Bullet to Low-Resource Languages Machine Translation?](https://md.research.elamur.ai/free/en/papers/song-2026-are-small-language-models-silver/), the authors create synthetic training material: strong teacher models translate text, and that output is used to train smaller models. The question is whether compact models trained this way can serve languages that lack large parallel corpora. [A Fine-Grained Linguistic Evaluation of Low-Resource Luxembourgish–English MT](https://md.research.elamur.ai/free/en/papers/rehlinger-2026-fine-grained-linguistic-evaluation-low/) takes the opposite angle. Instead of building a system, it builds a test. The Lux-MT-Test-Suite is presented as the first challenge set that probes Luxembourgish-to-English translation phenomenon by phenomenon, rather than through a single overall score. The third, [Contributing to Speech-to-Speech Translation for African Low-Resource Languages](https://md.research.elamur.ai/free/en/papers/ouedraogo-2026-contributing-speech-speech-translation-african/), studies the French–Mooré pair. Mooré is described there as the most widely used local language of Burkina Faso. This work also links the translation strand to the topic of [speech and spoken language](https://md.research.elamur.ai/free/en/topics/speech/).

A second strand concerns [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/), specifically Arabic. [The AraGenEval Shared Task on Arabic Authorship Style Transfer and AI Generated Text Detection](https://md.research.elamur.ai/free/en/papers/abudalfa-2025-arageneval-shared-task-arabic-authorship/) was held at ArabicNLP 2025, alongside EMNLP 2025. It asked participants to rewrite Arabic text in the style of particular authors and to tell human-written from machine-generated Arabic. The follow-on [AMIYA Shared Task: Arabic Modeling In Your Accent at VarDial 2026](https://md.research.elamur.ai/free/en/papers/robinson-2026-amiya-shared-task-arabic-modeling/) addresses a known weakness. Large language models handle Modern Standard Arabic reasonably well but drift back to it even when a user writes in a regional dialect. AMIYA sets out to measure and encourage dialect-faithful generation.

A third, smaller strand deals with [historical linguistics and language change](https://md.research.elamur.ai/free/en/topics/language-change/). [Semantic Change Characterization with LLMs using Rhetorics](https://md.research.elamur.ai/free/en/papers/sa-2026-semantic-change-characterization-llms-rhetorics/) starts from the observation that most research on lexical semantic change only asks whether a word's meaning has moved. It goes further and uses large language models, together with rhetorical categories, to describe how a meaning has shifted.

Finally, [evaluation and benchmarks](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/) and [reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) are represented by [Humanity's Last Exam](https://md.research.elamur.ai/free/en/papers/phan-2025-humanity-s-last-exam/). It responds to the problem that familiar benchmarks such as MMLU are close to saturation and no longer separate capable models well. Evaluation is in fact a thread through the whole profile. The benchmark, the two shared tasks and the Luxembourgish test suite are all ways of measuring what language systems can and cannot do.

### Leading scientists

All of the people with a page here are affiliated with the University of Luxembourg. They appear below in the order the country's people are presented, which runs from the most cited.

[Saad Ezzini](https://md.research.elamur.ai/free/en/authors/saad-ezzini/) comes first. Ezzini is a co-author of the AraGenEval shared task, and the Arabic language work is the clearest part of their contribution on this site. Ezzini's papers here have the widest reach among the country's researchers.

[Christoph Purschke](https://md.research.elamur.ai/free/en/authors/christoph-purschke/) follows. Purschke's work here sits within the university's research on language, and Purschke's papers rank high among its researchers in how often they are cited.

[Tegawende Bissyande](https://md.research.elamur.ai/free/en/authors/tegawende-bissyande/) is next. Bissyande's papers in our library belong to the same language-technology output of the university, which on this site spans translation, Arabic NLP and evaluation.

[Jacques Klein](https://md.research.elamur.ai/free/en/authors/jacques-klein/) is also part of this university group. Klein's papers here fall within that group's work on language systems.

[Salima Lamsiyah](https://md.research.elamur.ai/free/en/authors/salima-lamsiyah/) contributes papers that place Lamsiyah alongside the colleagues above in the university's work on natural language processing.

[Alistair Plum](https://md.research.elamur.ai/free/en/authors/alistair-plum/) has several papers here, all within the university's body of language research. That research ranges from shared tasks to studies of specific low-resource languages.

[Anne-Marie Lutgen](https://md.research.elamur.ai/free/en/authors/anne-marie-lutgen/) is another member of this circle at the university. Lutgen's papers in our library add to its language-focused output.

[Yewei Song](https://md.research.elamur.ai/free/en/authors/yewei-song/) is the lead author of the study on small language models for low-resource translation. Song's work here is therefore most directly tied to the idea of training compact translation models with synthetic data produced by larger ones. Through that paper, Song's work is also connected to the Luxembourg Institute of Science and Technology.

### Landmark works

The most cited work by far is Humanity's Last Exam (2025), with Long Phan, Alice Gatti and Ziwen Han among its many authors. It offers a hard, broad test of model capability at a time when older benchmarks no longer distinguish strong systems well. Much of the country's citation weight on this site rests on this one paper. Other works here use it as a reference point when they evaluate models on demanding reasoning questions.

The second most cited is the AraGenEval shared task (2025), led by Shadi Abudalfa with Saad Ezzini and Ahmed Abdelali among the co-authors. By pairing style transfer with the detection of AI-generated text, it turned two separate problems in Arabic NLP into a common evaluation that other teams could enter and compare against.

Among the 2026 works, the semantic change study and the small-language-model translation study have begun to draw citations here. The French–Mooré speech translation paper and the AMIYA shared task have each been taken up as well. The Luxembourgish–English test suite by Nils Rehlinger is the most recent kind of contribution: a single-author evaluation resource that other researchers working on Luxembourgish translation can build on.

### Labs and units

The country's research here runs through two institutions. The University of Luxembourg is the main one. It is the home of every person with a page and the affiliation of every work. The Luxembourg Institute of Science and Technology, in Belvaux, appears as a co-affiliation on two works, the semantic change study and the small-language-model translation study. There are no people with a page attached to it on this site. The full set of both is gathered on the country's [institutions](https://md.research.elamur.ai/free/en/countries/luxembourg/institutions/) view, and the single urban region on its [agglomerations](https://md.research.elamur.ai/free/en/countries/luxembourg/agglomerations/) view.

### Collaborations

The most visible domestic link is between the university and the research institute. Their shared papers deal with translation for low-resource languages and with tracking meaning change. On the semantic change paper, the university and the institute share a team that includes Jáder Martins Camboim de Sá, Jooyoung Lee and Marcos Da Silveira. On the translation study, Yewei Song works with Lujun Li and Cedric Lothritz.

Outside the country, the collaborations take several forms. Humanity's Last Exam is a broad multi-author project, with Luxembourg as one contributor among many. The Arabic shared tasks bring together organisers from the Arabic NLP community. Shadi Abudalfa and Ahmed Abdelali work on AraGenEval. Nathaniel R. Robinson, Shahd Abdelmoneim and Anjali Kantharuban are on AMIYA, which was held at VarDial 2026. The French–Mooré speech translation work is authored with Fayçal S. A. Ouedraogo, Maimouna Ouattara and Rodrique Kafando, and is concerned with a language of Burkina Faso. Taken together, these links tie research in Luxembourg to Arabic-speaking regions, to West Africa, and to large community benchmark efforts.

### Timeline

In 2025 the country's work in our library opens with two collaborative evaluation efforts. One is Humanity's Last Exam, a demanding benchmark for advanced models. The other is the AraGenEval shared task on Arabic authorship style transfer and machine-text detection. Both papers come from the University of Luxembourg and remain its most cited works here.

In 2026 the research spreads across specific languages and methods. Arabic work continues with the AMIYA task on dialectal generation. Translation for low-resource languages becomes the leading topic, through the study of small models trained on synthetic data, the French–Mooré speech translation study, and the Luxembourgish–English test suite. A study of semantic change with large language models adds a historical linguistics angle. In this period the Luxembourg Institute of Science and Technology also joins the university on shared papers.

## Where

## Agglomerations (1)

1. [**Luxembourg**2 institutions*285 citations*](https://md.research.elamur.ai/free/en/agglomerations/luxembourg/)

## Institutions (2)

### 🎓 Universities (1)

1. **[University of Luxembourg](https://md.research.elamur.ai/free/en/universities/university-of-luxembourg/)**University · [Luxembourg](https://md.research.elamur.ai/free/en/agglomerations/luxembourg/), Luxembourg*7 works · 279 citations*

### 🔬 Labs and research institutes (1)

1. **[Luxembourg Institute of Science and Technology](https://md.research.elamur.ai/free/en/labs/luxembourg-institute-of-science-and-technology/)**Research institute · Belvaux ([Luxembourg](https://md.research.elamur.ai/free/en/agglomerations/luxembourg/)), Luxembourg*2 works · 6 citations*

## People (8)

Authors with a page confirmed at its institutions, the most cited first.

1. [**Saad Ezzini**University of Luxembourg*6 papers here · 134 citations*](https://md.research.elamur.ai/free/en/authors/saad-ezzini/)
2. [**Christoph Purschke**University of Luxembourg*2 papers here · 79 citations*](https://md.research.elamur.ai/free/en/authors/christoph-purschke/)
3. [**Tegawende Bissyande**University of Luxembourg*3 papers here · 61 citations*](https://md.research.elamur.ai/free/en/authors/tegawende-bissyande/)
4. [**Jacques Klein**University of Luxembourg*2 papers here · 59 citations*](https://md.research.elamur.ai/free/en/authors/jacques-klein/)
5. [**Salima Lamsiyah**University of Luxembourg*2 papers here · 56 citations*](https://md.research.elamur.ai/free/en/authors/salima-lamsiyah/)
6. [**Alistair Plum**University of Luxembourg*3 papers here · 50 citations*](https://md.research.elamur.ai/free/en/authors/alistair-plum/)
7. [**Anne-Marie Lutgen**University of Luxembourg*2 papers here · 22 citations*](https://md.research.elamur.ai/free/en/authors/anne-marie-lutgen/)
8. [**Yewei Song**University of Luxembourg*2 papers here · 15 citations*](https://md.research.elamur.ai/free/en/authors/yewei-song/)

## Works in our library (7)

Works of its institutions, the most cited first.

1. **[Humanity's Last Exam](https://md.research.elamur.ai/free/en/papers/phan-2025-humanity-s-last-exam/) [Nature](https://md.research.elamur.ai/free/en/venues/nature/)**Long Phan, Alice Gatti, Ziwen Han, et al.*2025 · 235 citations*
2. **[The AraGenEval Shared Task on Arabic Authorship Style Transfer and AI Generated Text Detection](https://md.research.elamur.ai/free/en/papers/abudalfa-2025-arageneval-shared-task-arabic-authorship/) [Arabic Natural Language Processing Conference (ArabicNLP)](https://md.research.elamur.ai/free/en/venues/arabicnlp/)**Shadi Abudalfa, Saad Ezzini, Ahmed Abdelali, et al.*2025 · 36 citations*
3. **[Semantic Change Characterization with LLMs using Rhetorics](https://md.research.elamur.ai/free/en/papers/sa-2026-semantic-change-characterization-llms-rhetorics/) [Proceedings for the International Workshop on Computational Approaches to Language Change](https://md.research.elamur.ai/free/en/venues/lchange/)**Jáder Martins Camboim de Sá, Jooyoung Lee, Marcos Da Silveira, et al.*2026 · 3 citations*
4. **[Are Small Language Models the Silver Bullet to Low-Resource Languages Machine Translation?](https://md.research.elamur.ai/free/en/papers/song-2026-are-small-language-models-silver/) [Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT)](https://md.research.elamur.ai/free/en/venues/loresmt/)**Yewei Song, Lujun Li, Cedric Lothritz, et al.*2026 · 3 citations*
5. **[Contributing to Speech-to-Speech Translation for African Low-Resource Languages : Study of French-Mooré Pair](https://md.research.elamur.ai/free/en/papers/ouedraogo-2026-contributing-speech-speech-translation-african/) [Workshop on Language Models for Low-Resource Languages](https://md.research.elamur.ai/free/en/venues/loreslm/)**Fayçal S. A. Ouedraogo, Maimouna Ouattara, Rodrique Kafando, et al.*2026 · 1 citation*
6. **[AMIYA Shared Task: Arabic Modeling In Your Accent at VarDial 2026](https://md.research.elamur.ai/free/en/papers/robinson-2026-amiya-shared-task-arabic-modeling/) [Workshop on NLP for Similar Languages, Varieties and Dialects (VarDial)](https://md.research.elamur.ai/free/en/venues/vardial/)**Nathaniel R. Robinson, Shahd Abdelmoneim, Anjali Kantharuban, et al.*2026 · 1 citation*
7. **[A Fine-Grained Linguistic Evaluation of Low-Resource Luxembourgish–English MT](https://md.research.elamur.ai/free/en/papers/rehlinger-2026-fine-grained-linguistic-evaluation-low/) [Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT)](https://md.research.elamur.ai/free/en/venues/loresmt/)**Nils Rehlinger*2026*

## Fields, topics and concepts

Of its works in our library, with the number of works on each.

📚 Shelves 7 The field of science this work belongs to (OpenAlex fields).[Arts and Humanities · 7 works](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · [Computer Science · 7 works](https://md.research.elamur.ai/free/en/shelves/computer-science/) · [Artificial Intelligence · 7 works](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · [Language and Linguistics · 7 works](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · [Computer Vision and Pattern Recognition · 1 work](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · [Mathematics · 1 work](https://md.research.elamur.ai/free/en/shelves/mathematics/) · [Signal Processing · 1 work](https://md.research.elamur.ai/free/en/shelves/computer-science/signal-processing/)

🧭 Topics 9 Research subjects and languages the work is about.[Semitic languages · 2 works](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · [Low-resource languages · 3 works](https://md.research.elamur.ai/free/en/topics/low-resource-languages/) · [Machine translation · 3 works](https://md.research.elamur.ai/free/en/topics/machine-translation/) · [Historical linguistics & language change · 1 work](https://md.research.elamur.ai/free/en/topics/language-change/) · [Reasoning · 1 work](https://md.research.elamur.ai/free/en/topics/reasoning/) · [French · 1 work](https://md.research.elamur.ai/free/en/topics/french/) · [Speech & spoken language · 1 work](https://md.research.elamur.ai/free/en/topics/speech/) · [English · 1 work](https://md.research.elamur.ai/free/en/topics/english/) · [Evaluation & benchmarks · 1 work](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/)

## Works by year

### 2020s

[**2026** 5 works Works of Luxembourg from 2026](https://md.research.elamur.ai/free/en/countries/luxembourg/year/2026/) · [**2025** 2 works Works of Luxembourg from 2025](https://md.research.elamur.ai/free/en/countries/luxembourg/year/2025/)

- [← back to](https://md.research.elamur.ai/free/en/countries/) — countries

Updated 24 October 2026