Jeremie Pantin

Université de Caen Normandie · GREYC (IMAGE team) · jeremie.pantin [at] unicaen.fr

I am currently working as an assistant professor (CTER) for Université de Caen Normandie, and doing research at the GREYC laboratory.

After completing my Master of Science in Computer Science at Rouen University, I worked as a Data Scientist for a French startup based in Paris, specializing in Natural Language Processing for marketing projects (consumer journey, ...). Subsequently, I pursued a Ph.D. in Artificial Intelligence at Sorbonne Université in Paris, within the LFI team of LIP6 lab. My research, supervised by Christophe Marsala, focused on Machine Learning and Anomaly Detection for textual data. In 2024 I became assistant professor (ATER) at Sorbonne University, before joining University of Caen Normandy.

Experience

CTER

Université de Caen Normandie - GREYC

Teaching computer science at Caen University and conducting research at GREYC laboratory. The research mainly focuses to Anomaly Detection, NLP and Explainable AI. With major application to AI for good.

2024 - now

ATER

Sorbonne Université - LIP6

Section CNU 27

Teaching computer science at Sorbonne Université and conducting research at LIP6 laboratory. The research mainly focuses to Anomaly Detection, NLP and Explainable AI.

2023 - 2024

Ph.D. Student

Sorbonne Université - LIP6
Title: Characterization and detection of semantic textual outliers
Funding:
  • University contract (2020-2022)
  • ATER (2022-2023)
Publicly defended on September 11th, 2023
Keywords: Machine Learning, Anomaly Detection, NLP, Autoencoders, Abstractive Summarization, Weak Signal Detection

2020-2023

(Lead) Data Scientist

TheDataStrategy

Applied artificial intelligence to marketing projects and served as a data science consultant for various companies. Led data science teams and supervised junior data scientists, software engineer interns, and data engineer interns. Oversaw data science projects for multiple companies:

  • Weak signal detection on social media for marketing strategy (Text, Graph, Web)
  • Weak signal detection for financial application (Time series, Text)
  • Image segmentation for document analysis (Image-to-Text)
  • Image segmentation on satelite images (Image)
  • Text summarization for social media analysis (Text, Graph)

2017 - 2020

Research

My research in artificial intelligence centers on machine learning and natural language processing (NLP), with a specific focus on anomaly detection. One of the first contribution was the identification of a taxonomy for textual anomalies, which display a different view of results from state of the art approaches. Subsequently, I introduced a novel approach using autoencoders to detect anomalies in text. Recently, I've integrated my thesis work on explaining anomaly detection models with recent advancements in eXplainable AI (XAI) and representation learning in Large Language Models. This research addresses challenges in state-of-the-art models, notably their sensitivity to counterfactuals and out-of-distribution data. Using anomaly analysis, I've demonstrated an improvement in the quality of abstractive summarization with unsupervised approaches.

Keywords: Machine Learning, Unsupervised Learning, Anomaly Detection, Natural Language Processing, Robust Generative AI.

Conferences

2026
  • A. Youcef Khodja, J. Pantin, G. Dias, & F. Maurel. (2026). Anomaly Detection using Source-Reversed Flow Matching. Proceedings of the 31st Conference on Empirical Methods in Natural Language Processing. Just accepted.
  • R. Lemaire, A. Kaibaldiyev, E. Mariette, D. Viglieri, J. Pantin, A. Lechervy, , F. Maurel, G. Dias, G. Blaizot, V. Agin, N. Poirel, E. Bui, H. Platel, D. Vivien, & Y. Chahir. (2026). MONET: Museum Observatory of Neuro-Affective Eye-Tracking. Proceedings of the 34th ACM International Conference on Multimedia. Just accepted.
  • R. Pinninti, L. C. Routhu, D. Bandyopadhyay, G. Dias, A. Lechervy, J. Pantin, M. Hasanuzzaman, & A. Ekbal. (2026). Corrective Feedback-Driven Humor Generation for Tiny Language Models. Proceedings of the 5th AACL and the 15th IJCNLP. Just accepted.
  • R. Lemaire, J. Pantin, A. Lechervy, A. Kaibaldiyev, F. Maurel, G. Dias, & Y. Chahir. (2026). Visually Informed Text Representations for Visual Contextual Classification of Arts. International Conference on Pattern Recognition (pp. 285–300). [doi] [pdf].
2025
  • J. Pantin, & C. Marsala. (2025). A Robust Autoencoder Ensemble-Based Approach for Anomaly Detection in Text. Proceedings of the 29th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (pp. 1986-1995). [doi] [pdf]
  • F. Ledoyen, G. Dias, J. Pantin, A. Lechervy, F. Maurel, & Y. Chahir. (2025). Facilitating Cognitive Accessibility with LLMs: A Multi-Task Approach to Easy-to-Read Text Generation. Proceedings of the 30th Conference on Empirical Methods in Natural Language Processing (pp. 11782-11808). [doi] [pdf]
  • F. Ledoyen, G. Dias, A. Lechervy, J. Pantin, F. Maurel, Y. Chahir, E. Gouzonnat, M. Berthelot, S. Moravac, A. Altinier, & A. Khairalla. (2025). Inclusive Easy-to-Read Text Generation for Individuals with Cognitive Impairments. Proceedings of the 28th European Conference on Artificial Intelligence (pp. 3503-3510). [doi] [pdf]
  • R. Lemaire, A. Kaibaldiyev, E. Mariette, D. Viglieri, A. Lechervy, F. Maurel, G. Dias, J. Pantin, G. Blaizot, V. Agin, N. Poirel, E. Bui, H. Platel, D. Vivien, & Y. Chahir. (2025). Inclusive Easy-to-Read Text Generation for Individuals with Cognitive Impairments. Proceedings of the 33rd ACM International Conference on Multimedia (pp. 13582-13584). [doi] [pdf]
  • A. Kaibaldiyev, J. Pantin, A. Lechervy, F. Maurel, Y. Chahir, & G. Dias. (2025). UNETRSAL: Saliency Prediction with Hybrid Transformer-Based Architecture. Proceedings of the 22nd International Conference on Advanced Concepts for Intelligent Vision Systems (pp. 389-400). [doi] [pdf]
  • D.E. Merzougui, N. Tete, F. Maurel, G. Dias, M. Hasanuzzaman, A. Bournonville, E. Madelaine, T. Berthelin Le Tellier, F. Ledoyen, L. Poutrain-Lejeune, F. Rioult, & J. Pantin. (2025). Concurrent Speech and Auditory Tag Clouds for Non-Visual Web Interaction. Proceedings of Interspeech 2025 (pp. 4963-4964). [pdf]
  • D.E. Merzougui, G. Dias, J. Pantin, & F. Maurel (2025). Evaluating Large Language Models for Depression Symptom Estimation. Proceedings of the 23rd International Conference on Artificial Intelligence in Medicine (pp. 272-276). [doi] [pdf]
2024
  • J. Pantin, & C. Marsala. (2024). Détection d'anomalies textuelles par ensemble d'autoencodeurs robustes. Conférence sur l'Extraction et la Gestion des Connaissances (pp. 319-326). [pdf]

Workshops

2026
  • J. Lovon-Melgarejo, J. Pantin, & G. Dias. (2026). GREYC at CRF Filling 2026: Rewrite Before You Extract - Rewriting Clinical Notes for Automated CRF. Proceedings of the Third Workshop on Patient-Oriented Language Processing (CL4Health) @ LREC 2026 (pp. 384–389). [doi][pdf]
2022
  • J. Pantin, M.-J. Lesot, & C. Marsala. (2022). Analyse de Données Aberrantes pour le Texte: Taxonomie et Etude Expérimentale. Conférence sur l'Extraction et la Gestion des Connaissances, Atelier TextMine' (pp. 15-26). [pdf]

Teaching

My experience as a teaching assistant, acquired during my Ph.D. and my ATER appointment, includes teaching both theoretical and practical computer science to first/second year undergraduate students at Sorbonne Université. More recently, as an assistant professor, I expanded my teaching activities to Master's-level courses (both M1 and M2). Starting in September 2025, I will serve as the responsible lecturer for an advanced engineering course in machine learning at Université Caen Normandie.

Teaching Code Responsibility Level University Year
Advanced engineering for machine learning SMINF3E9 CM/TP (Resp. Lecturer) M2 Université Caen Normandie 2025/2026
Web development (PHP) SINFL4A1 TP L2 Université Caen Normandie 2025/2026
Natural Language Processing (NLP) SMINF2E6 CM M1 Université Caen Normandie 2025/2026
Advanced NLP SMINF3E1 CM/TP M2 Université Caen Normandie 2024/2026
Introduction to NLP SMINF1E1 TD M1 Université Caen Normandie 2024/2026
Project SMINFO1D Project M1 Université Caen Normandie 2024/2025
Probability & Statistics SMINF1E1 TD/TP M1 Université Caen Normandie 2024
Web development (HTML/CSS) YSINL1C1 TP L1 Université Caen Normandie 2024
Data structure LU2IN006 TD/TP L2 Sorbonne Université 2023
Databases LU2IN009 TP L2 Sorbonne Université 2023
Scientific programming 2 (C) LU1IN002 TP L1 Sorbonne Université 2023-2024
Scientific programming 1 (Python) LU1IN011 TD/TP/Project L1 Sorbonne Université 2022-2023
Object-oriented programming (JAVA) LU2IN002 TP/TP L2 Sorbonne Université 2019-2022

Ph.D. Students

  • (2025-now) A. Youcef Khodja - Directed and co-supervised with G. Dias and F. Maurel

Supervising

In a research context, I have supervised two data scientist interns and one data engineer apprentice.

2019 - Data engineer

Supervision of a Computer Science Master's apprentice in micro-service development and continuous integration. The context of the apprenticeship revolves around a new challenge, which involves deploying a new distributed platform accommodating numerous artificial intelligence models. A second aspect also includes the integration of the Kotlin language for platform development.

2018 - Machine learning scientist

Supervision of a final year master's student intern. The supervision involved defining tasks in machine learning and natural language processing (NLP), as well as overseeing the progress of a research project during the 6-month internship. The research project focused on line segmentation from Curriculum Vitae images and extracting information through character recognition. I was also responsible of reviewing the final thesis.

2018 - Data scientist

Supervision of a final year master's student intern on the fundamental concepts of data science: web development (front-end and back-end), deployment of databases and microservices in a distributed infrastructure, and artificial intelligence research project. The research project focused on object recognition from satellite images. Additionally, I was responsible for reviewing the internship report and providing advices on best practices in software development and research best-practices.