Research

The Les femmes dans les métiers du cinéma français (2025-2026): is a collaborative research project, funded by ArTeC that aims to bring visibility to women professionals in the history of french cinema. I'm leading the technological development on behalf of LIASD. The project’s goal is to identify, count, and document these women's careers by mining archival records from major cultural institutions. In collaboration with the research lab of EPITA, we apply artificial intelligence techniques to process historical professional directories and cinema-related press preserved at the Bibliothèque nationale de France (BnF) and the Cinémathèque française.

The CLexIC project (2023) : The crowdfunding ecosystem offers a particularly rich field for study due to the diversity of proposed projects and the phenomenon of collective funding from the 'crowd'. Descriptions of activities and projects on crowdfunding platforms differ from product and service descriptions found on corporate websites. In the CLexIC project, our goal is to identify and analyze the terms that denote innovation, as well as the context of their usage. New terms will be contrasted with those already established by experts and through machine learning on corpora, from previous project MALANTIN.

A PhD project with NOVAGEN (April 2019-2022). I supervise the research of Maroua Boudabous in acquisition of web data, dataset creation, aspect recognition and prediction.

The Litte_Bot project (2019-2021): is a collaborative project, funded by ArTeC around conversational agents, chatbots, interacting with characters from 17th century theater. It is a dramaturgy for interacting with the character of Don Juan, as a part of the 400 th anniversary of Molière's birth in 2022. My work is about the construction of the dataset(s) and use of different methods and models to generating dialogues between the human and the machine. The bot gives the cue as if it embodied a character from Molière’s theater.

The MALANTIN project (MAchine Learning for the Automated textual analysis of Non-Technological INnnovation) (2018-2020) : This project brings together computer scientists, linguists and economists, around the problem of collecting and exploring multisource data on the theme of the recognition of non-technological innovation. The MALANTIN project is an exploratory research project between the laboratories: LISIS (ESIEE, France), LIGM (Université Gustave Eiffel, France) and LIASD (University Paris-8, France). My work in this project is supervising the technological aspects such as crawling (research of the companies responding to the criteria of innovation), scraping (extraction of textual data from the web), parsing (analyzing and annotating multilingual data for the dataset, curation/preprocessing), and implementing machnine learning (mostly deep) models.

2025 :

A. Pappa Corpus textuels : enjeux et défis. À paraître : 1024 bulletin de la Société Informatique de France (SIF) numéro 25 juin 2025.

R. Kyriakoglou, A. Pappa, J. He, A. Schoen, P. Laurens, M. Vartampetian, Ph. Laredo, T. Kyriacopoulou. Développement automatique de lexiques pour les concepts émergents : une exploration méthodologique. 17es Journées internationales d'Analyse statistique des Données Textuelles JADT 2024. Article

M. Boudabous and A. Pappa. Explicit aspect annotation via transfer and active learning. 27th International Conference on Knowledge Based and Intelligent Information and Engineering Sytems, Procedia Computer Science, 225:1124–1133, 2023. Article

A. Pappa, S. Szoniecky, R. Berenguer, J. Huthwohl, C. Quach and A. Laborderie. Designing a database for a classical theatrical chatbot. In 52nd LIBER Annual Conference, pages 153–155, 2023. Presentation

T. Cazenave, G. Grosjean, B. Rozière et A. Pappa. Moliaire, a theatrical agent which speaks like Molière’s characters. In 25th Generative Art Conference (GA), pages 286–294, ISBN 978–88–96610–44–2, Rome, Italy, 2022. Article

M. Boudabous et A. Pappa. Apprentissage actif pour l’extraction des aspects explicites : application à des avis non annotés en français. In Journées Jointes des Groupements de Recherche Linguistique Informatique, Formelle et de Terrain (LIFT) et Traitement Automatique des Langues (TAL), pages 20–28, Marseille, France, 2022. Article

A. Pappa, T. Cazenave, G. Grosjean and B. Rozière. Un chatbot qui donne la réplique ! Atelier Affects, Compagnons Artificiels et Interactions (ACAI), AFIA. École des Mines de Saint-Etienne, du 27 juin au 1er juillet 2022. Poster

G. Grosjean, A. Pappa, B. Rozière and T. Cazenave. Dialogue avec Molière. 29e conférence sur le Traitement Automatique des Langues Naturelles, TALN-RECITAL-ATALA. Avignon, du 27 juin au 1er juillet 2022. Poster

A. Pappa. Conférence invitée, Litte_Bot : a theatrical dialogue system using Seq2Seq model. Journée AI4LAM Community en distanciel, sur le thème Content Generative Models and GLAM. 19 avril 2022. Invited talk

A. Pappa. Litte_Bot : Le bot qui donne la réplique dans le style Molière. Journée JAII-2022, Accès interactif à l’information (AFIA-THL / ARIA / GDR TAL), page 22, Paris, 5 mai 2022. Poster

M. Boudabous and A. Pappa. WebT-IDC : A WebTool for Intelligent Dataset Creation, a use case for forums and blogs. In 25th International Conference Knowledge-Based and Intelligent Information & Engineering Systems, pages 1051–1060, vol 192, Elsevier, 2021. Article

A. Pappa, R. Berenguer and A. Laborderie. Litte_Bot, un agent conversationnel : une dramaturgie pour dialoguer avec Dom Juan. Journées de pré-conférence : Retours d’expériences pratiques et techniques. Les Futurs Fantastiques 2021 #FF21, Presentation/Demo. Université Paris Saclay, le 8 décembre 2021. Presentation

A. Pappa, R. Berenguer and A. Laborderie. Litte_Bot, un agent conversationnel : une dramaturgie pour dialoguer avec Dom Juan. Journée de la Conférence : Les défis de l’IA dans les bibliothèques, archives et musée. BnF, le 10 décembre 2021. Poster

The SALTY (2009-2012) project (Self-Adaptive very Large disTributed sYstems), funded by the ANR under contract ANR-09-SEGI-012.