Marco Naguib

Marco Naguib

Auto-generated profile, based on publicly available information

Marco Naguib

Université Paris-Saclay, CNRS, Laboratoire Interdisciplinaire des Sciences du Numérique, Orsay, France

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 September 2024 in Journal of Medical Internet Research
Verified
Journal Article
Open access
JZJamil Zaghir
Jamil Zaghir
MNMarco Naguib
Marco Naguib
MBMina Bjelogrlic
Mina Bjelogrlic
ANAurélie Névéol
Aurélie Névéol
XTXavier Tannier
Xavier Tannier
CLChristian Lovis
Christian Lovis
Background: Prompt engineering, focusing on crafting effective prompts to large language models (LLMs), has garnered attention for its capabilities at harnessing the potential of LLMs. This is even more crucial in the medical domain due to its specialized terminology and language technicity. Clinical natural language processing applications must navigate complex language and ensure privacy compliance. Prompt engineering offers a novel approach by designing tailored prompts to guide models in exploiting clinically relevant information from complex medical texts. Despite its promise, the efficacy of prompt engineering in the medical domain remains to be fully explored. Objective: The aim of the study is to review research efforts and technical approaches in prompt engineering for medical applications as well as provide an overview of opportunities and challenges for clinical practice. Methods: Databases indexing the fields of medicine, computer science, and medical informatics were queried in order to identify relevant published papers. Since prompt engineering is an emerging field, preprint databases were also considered. Multiple data were extracted, such as the prompt paradigm, the involved LLMs, the languages of the study, the domain of the topic, the baselines, and several learning, design, and architecture strategies specific to prompt engineering. We include studies that apply prompt engineering–based methods to the medical domain, published between 2022 and 2024, and covering multiple prompt paradigms such as prompt learning (PL), prompt tuning (PT), and prompt design (PD). Results: We included 114 recent prompt engineering studies. Among the 3 prompt paradigms, we have observed that PD is the most prevalent (78 papers). In 12 papers, PD, PL, and PT terms were used interchangeably. While ChatGPT is the most commonly used LLM, we have identified 7 studies using this LLM on a sensitive clinical data set. Chain-of-thought, present in 17 studies, emerges as the most frequent PD technique. While PL and PT papers typically provide a baseline for evaluating prompt-based approaches, 61% (48/78) of the PD studies do not report any nonprompt-related baseline. Finally, we individually examine each of the key prompt engineering–specific information reported across papers and find that many studies neglect to explicitly mention them, posing a challenge for advancing prompt engineering research. Conclusions: In addition to reporting on trends and the scientific landscape of prompt engineering, we provide reporting guidelines for future studies to help advance research in the medical field. We also disclose tables and figures summarizing medical prompt engineering papers available and hope that future contributions will leverage these existing works to better advance the field.

14 May 2024 in JMIR Preprints
Verified
Preprint
Open access
JZJamil Zaghir
Jamil Zaghir
MNMarco Naguib
Marco Naguib
MBMina Bjelogrlic
Mina Bjelogrlic
ANAurélie Névéol
Aurélie Névéol
XTXavier Tannier
Xavier Tannier
CLChristian Lovis
Christian Lovis
BACKGROUND Prompt engineering, focusing on crafting effective prompts to large language models (LLMs), has garnered attention for its capabilities at harnessing the potential of LLMs. This is even more crucial in the medical domain due to its specialized terminology and language technicity. Clinical natural language processing applications must navigate complex language and ensure privacy compliance. Prompt engineering offers a novel approach by designing tailored prompts to guide models in exploiting clinically relevant information from complex medical texts. Despite its promise, the efficacy of prompt engineering in the medical domain remains to be fully explored. OBJECTIVE The aim of the study is to review research efforts and technical approaches in prompt engineering for medical applications as well as provide an overview of opportunities and challenges for clinical practice. METHODS Databases indexing the fields of medicine, computer science, and medical informatics were queried in order to identify relevant published papers. Since prompt engineering is an emerging field, preprint databases were also considered. Multiple data were extracted, such as the prompt paradigm, the involved LLMs, the languages of the study, the domain of the topic, the baselines, and several learning, design, and architecture strategies specific to prompt engineering. We include studies that apply prompt engineering–based methods to the medical domain, published between 2022 and 2024, and covering multiple prompt paradigms such as prompt learning (PL), prompt tuning (PT), and prompt design (PD). RESULTS We included 114 recent prompt engineering studies. Among the 3 prompt paradigms, we have observed that PD is the most prevalent (78 papers). In 12 papers, PD, PL, and PT terms were used interchangeably. While ChatGPT is the most commonly used LLM, we have identified 7 studies using this LLM on a sensitive clinical data set. Chain-of-thought, present in 17 studies, emerges as the most frequent PD technique. While PL and PT papers typically provide a baseline for evaluating prompt-based approaches, 61% (48/78) of the PD studies do not report any nonprompt-related baseline. Finally, we individually examine each of the key prompt engineering–specific information reported across papers and find that many studies neglect to explicitly mention them, posing a challenge for advancing prompt engineering research. CONCLUSIONS In addition to reporting on trends and the scientific landscape of prompt engineering, we provide reporting guidelines for future studies to help advance research in the medical field. We also disclose tables and figures summarizing medical prompt engineering papers available and hope that future contributions will leverage these existing works to better advance the field.

9 April 2026
Verified
Thesis
Closed access
MNMarco Naguib
Marco Naguib
Les systèmes de santé modernes produisent quotidiennement d'importants volumes de données cliniques non structurées (des textes comme les comptes rendus d'hospitalisation, lettres de sortie, résultats d'examens), qui sont riches en informations essentielles pour le suivi des patients, la recherche médicale et la santé publique. Toutefois, la diversité linguistique, la complexité sémantique et les contraintes de confidentialité limitent fortement l'exploitation automatique de ces textes, en particulier en langue française. L'extraction d'information clinique à partir de ces documents demeure ainsi un défi majeur, surtout dans un contexte où les corpus annotés sont rares et les protocoles d'évaluation établis ne sont pas toujours applicables et standardisés.Cette thèse se positionne à l'intersection de l'informatique médicale et du traitement automatique des langues (TAL), et cible spécifiquement les documents cliniques francophones. Elle explore la reconnaissance d'entités nommées (REN) et le résumé automatique de dossier patient en contexte d'urgence.Afin de répondre aux défis méthodologiques, techniques et évaluatifs propres à l'extraction d'information en français, la thèse s'articule autour de quatre axes de recherche : le manque de corpus annotés, l'adaptation des modèles de langues généralistes au domaine clinique, l'évaluation rigoureuse dans des contextes hospitaliers réels, et l'identification des scénarios d'usage adaptés aux grands modèles de langues.Les résultats obtenus montrent que, pour la reconnaissance d'entités cliniques, dans un contexte où il y a peu de corpus annotés disponibles en français, les modèles de langues masqués affinés surpassent nettement les grands modèles génératifs, aussi bien en termes de performance qu'en impact environnemental. L'apprentissage actif appliqué à la sélection d'exemples permet, en outre, de réduire considérablement le coût d'annotation sans perte significative de performance. Concernant le résumé de dossier patient, une approche RAG associant récupération symbolique des passages pertinents et génération automatique par LLM permet d'obtenir des synthèses à la fois cliniquement pertinentes et robustes, contribuant ainsi à une amélioration de l'accessibilité de l'information critique pour la prise en charge rapide des patients. Ces contributions sont évaluées à l'aide de protocoles reproductibles, intégrant des métriques classiques, des analyses humaines qualitatives, et la mesure de l'impact environnemental.Au-delà des avancées méthodologiques, cette thèse invite à une intégration critique, transparente et contextualisée du TAL dans le domaine médical. Elle souligne les risques d'une confiance excessive envers des systèmes opaques dits « IA », l'importance de la confidentialité des données, les coûts environnementaux et sociétaux du TAL à grande échelle, et la nécessité d'une gouvernance et d'une documentation rigoureuses pour les applications en santé. Les travaux réalisés ouvrent la voie vers une extraction d'information clinique francophone plus fiable, durable et équitable, avec des perspectives étendues aux langues et contextes dans lesquels peu de corpus annotés sont disponibles.