Elena V. Epure
Tagline:Senior NLP Research Scientist at Deezer
Paris, France
About Me
I am a Senior NLP Research Scientist at Deezer. My current research focuses on creative media, building on my previous experience with social media and news content. My work applies NLP fundamentals to enhance the understanding, personalization, and contextualization of creative media. I also explore topics related to conversational and natural language recommendation and search. Drawing on a multidisciplinary educational background in computer science, digital humanities, and business, along with professional experience in research, software engineering, consultancy, and teaching, I can bring a cross-domain perspective to developing user-centered creative media systems.
Work Experiences
Senior NLP Research Scientist
from: 2018, until: presentOrganization:DeezerLocation:Paris, France
Research Consultant, part-time
from: 2017, until: 2017Organization:Panthéon-Sorbonne UniversityLocation:Paris, France
Software Engineer
from: 2014, until: 2014Organization:Trimble (former Stabiplan)Location:The Netherlands
Software Engineer, part-time
from: 2011, until: 2013Organization:42windmillsLocation:The Netherlands
Education
PhD
from: 2014, until: 2018Field of study:Computer ScienceSchool:Panthéon-Sorbonne UniversityLocation:Paris, France
Description-
Thesis: "Automatically modeling conversations as processes of interrelated speech intentions"
-
Intersected areas: natural language processing, pragmatics, conversation analysis, text and process mining
-
Side project on user modeling for news recommender systems
-
MSc
from: 2011, until: 2013Field of study:Business InformaticsSchool:Utrecht UniversityLocation:Utrecht, the Netherlands
Description- Thesis: "Intention-mining: a solution to process participant support in process-aware information systems"
BSc
from: 2007, until: 2011Field of study:Computer Science and EngineeringSchool:University Politehnica of BucharestLocation:Bucharest, Romania
Description- Thesis: "Evaluating brokers performance in a virtual stock market" (a multi-agent system with integrated reputation model and neural networks for trading)
Supervisions
- JA
Joanne Affolter
Research internship: "Musical versions identification from lyrics representations"
date: 2025 - 2025Degree: Master's Degree .
- AB
Abigail Berthe-Pardo
Research internship: "Persona extraction to support audiobook generation"
date: 2025 - 2025Degree: Master's Degree .
- MF
Markus Frohmann
Research internship: "Synthetic lyrics detection from audio"
date: 2024 - 2025Degree: Master's Degree .
- YL
Yanis Labrak
Research internship: "Synthetic lyrics detection from text"
date: 2024 - 2024Degree: Doctoral Degree .
- GM
Gaspard Michel
PhD thesis: "Automatic book understanding for an enhanced audiobook narration"
date: 2023 - presentDegree: Doctoral Degree .
- ND
Noé Durandard
Research internship: "Book natural language understanding"
date: 2022 - 2022Degree: Master's Degree .
- GA
Gamar Azuaje
Research internship: "Multimodal album art generation"
date: 2021 - 2022Degree: Master's Degree .
- GG
Giovanni Gabbolini
Research internship: "Automatic playlist captioning"
date: 2021 - 2022Degree: Doctoral Degree .
- VM
Vipul Mishra
Research internship: "A genre exploration of the emotions of music listeners on Reddit"
date: 2021 - 2021Degree: Master's Degree .
- FV
Francisco Valero
Research internship: "Podcast topic modelling"
date: 2020 - 2021Degree: Master's Degree .
Teachings
Introduction to Machine Learning
From: 2021, Until: 2022
Organization:Panthéon-Sorbonne UniversityField:Computer Science
Description:-
MSc. level course taught in English
-
Roles: Course co-creator, Lecturer, Seminar supervision
-
Process Mining
From: 2018, Until: 2020
Organization:Panthéon-Sorbonne UniversityField:Computer Science
Description:- Roles: Lecturer, Seminar supervision
Research methods
From: 2018, Until: 2020
Organization:Panthéon-Sorbonne UniversityField:Computer Science
Description:-
MSc. level course taught in French
-
Roles: Invited Lecturer
-
C Language Introduction
From: 2015, Until: 2015
Organization:Panthéon-Sorbonne UniversityField:Computer Science
Description:-
BSc. level seminar held in French
-
Roles: Seminar supervision
-
Microsoft .NET Programming
From: 2009, Until: 2011
Organization:Microsoft IT Academy UPBField:Computer Science
Description:-
Students and industry professionals, taught in Romanian
-
Roles: Course co-creator, Trainer including seminar supervision
-
Conference Contributions
ACM Conference on Recommender Systems (RecSys)
From: 2025
Women in RecSys co-chair
ACL Rolling Review (ARR)
From: 2022
To: present
Reviewer
North American Chapter of the Association for Computational Linguistics - Human Language Technologies Conference (NAACL-HLT)
From: 2021
Reviewer
Workshop on NLP for Music and Audio (NLP4MusA)
From: 2020
To: 2024
Co-organiser
Workshop on NLP for Music and Audio (NLP4MusA)
From: 2020
To: 2024
Reviewer
International Society for Music Information Retrieval Conference (ISMIR)
From: 2020
To: 2022
Reviewer
Empirical Methods in Natural Language Processing Conference (EMNLP)
From: 2020
2020
Workshop on Information Systems Engineering for Smarter Life (ISESL)
From: 2020
Reviewer
ACM Conference on Recommender Systems (RecSys)
From: 2019
To: 2020
Reviewer
IEEE International Conference on Research Challenges in Information Science (RCIS)
From: 2018
To: 2020
Posters and demos reviewer
Talks
Modeling User Preferences for Music Recommendations with LLMs: Opportunities and Limitations
Date: Jun 2025
Event name: Mixtapes Laboratory, CNRS .Location: Paris, France .
Modeling User Preferences for Music Recommendations with LLMs: Opportunities and Limitations
Date: May 2025
Event name: Loria Laboratory, CNRS .Location: Nancy, France .
Modeling User Preferences for Music Recommendations with LLMs: Opportunities and Limitations
Date: Feb 2025
Event name: MédiaLab, SciencesPo .Location: Paris, France .
NLP Meets Creative Media: Research Applications, Challenges and Opportunities
Date: Oct 2024
Event name: Idiap Research Institute .Location: Martigny, Switzerland .
Leveraging Book NLU for Enhanced AI Narration
Date: Oct 2024
Event name: EPFL .Location: Lausanne, France .
Honors & Awards
Outstanding Paper Award
date: 2023-05-03Issuer:European Chapter of the ACL (EACL)
Description:For the article “A Human Subject Study of Named Entity Recognition (NER) in Conversational Music Recommendation Queries”
Short-listed for Best Reviewer Award
date: 2019-09-17Issuer:ACM Recommender Systems Conference (RecSys)
Prix de thèse
date: 2019-06-13Issuer:Association INFORSID (INFormatique des ORganisations et Systèmes d’Information et de Décision)
Prix de thèse
date: 2019-04-12Issuer:Commission de la recherche de l'Université Paris 1 Panthéon-Sorbonne
Description:PhD prize for high academic and scientific work
ACM womENcourage Attendance Scholarship
date: 2016-09-12Issuer:Google
Doctoral research grant
date: 2014-10-01Issuer:Panthéon-Sorbonne University
Merit Scholarship for Outstanding Academic Performance
date: 2013-10-01Issuer:University Politehnica of Bucharest
Publications
Double Entendre: Robust Audio-Based AI-Generated Lyrics Detection via Multi-View Fusion
Conference PaperPublisher:Findings of the Association for Computational Linguistics: ACL 2025Date:2025Authors:Markus FrohmannGabriel Meseguer-BrocalMarkus SchedlElena V. EpureDescription:The rapid advancement of AI-based music generation tools is revolutionizing the music industry but also posing challenges to artists, copyright holders, and providers alike. This necessitates reliable methods for detecting such AI-generated content. However, existing detectors, relying on either audio or lyrics, face key practical limitations: audio-based detectors fail to generalize to new or unseen generators and are vulnerable to audio perturbations; lyrics-based methods require cleanly formatted and accurate lyrics, unavailable in practice. To overcome these limitations, we propose a novel, practically grounded approach: a multimodal, modular late-fusion pipeline that combines automatically transcribed sung lyrics and speech features capturing lyrics related information within the audio. By relying on lyrical aspects directly from audio, our method enhances robustness, mitigates susceptibility to low-level artifacts, and enables practical applicability. Experiments show that our method, DE-detect, outperforms existing lyrics-based detectors while also being more robust to audio perturbations. Thus, it offers an effective, robust solution for detecting AI-generated music in real-world scenarios. Our code is available at https://github.com/deezer/robust-AI-lyrics-detection.Just Ask for Music (JAM): Multimodal and Personalized Natural Language Music Recommendation
Conference PaperPublisher:Proceedings of the 19th ACM Conference on Recommender SystemsDate:2025Authors:Alessandro B. MelchiorreMarkus SchedlEvaluating LLMs for Quotation Attribution in Literary Texts: A Case Study of LLaMa3
Conference PaperPublisher:Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)Date:2025Authors:Gaspard MichelElena V. EpureRomain HennequinChristophe CerisaraDescription:Large Language Models (LLMs) have shown promising results in a variety of literary tasks, often using complex memorized details of narration and fictional characters. In this work, we evaluate the ability of Llama-3 at attributing utterances of direct-speech to their speaker in novels. The LLM shows impressive results on a corpus of 28 novels, surpassing published results with ChatGPT and encoder-based baselines by a large margin. We then validate these results by assessing the impact of book memorization and annotation contamination.We found that these types of memorization do not explain the large performance gain, making Llama-3 the new state-of-the-art for quotation attribution in English literature. We release publicly our code and data.Biases in LLM-Generated Musical Taste Profiles for Recommendation
Conference PaperPublisher:Proceedings of the 19th ACM Conference on Recommender SystemsDate:2025Authors:Harin Lee Bruno SguerraManuel MoussallamSynthetic Lyrics Detection Across Languages and Genres
Conference PaperPublisher:Proceedings of the 5th Workshop on Trustworthy NLP (TrustNLP 2025)Date:2025Authors:Yanis LabrakMarkus FrohmannGabriel Meseguer-BrocalElena V Epure