Publications by OLST members

@inproceedings{petrov_alf_2025,

title = {ALF: A Fine-Grained French Analogical Dataset for Evaluating Lexical Knowledge of Large Language Models},

author = {Alexander Petrov and Antoine Venant and François Lareau and Yves Lepage and Philippe Langlais},

year  = {2025},

date = {2025-10-01},

booktitle = {Proceedings of the 28th European Conference on Artificial Intelligence},

address = {Bologna, Italia},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{li_assessing_2025,

title = {Assessing LLMs' Understanding of Structural Contrasts in the Lexicon},

author = {Shuxu Li and Antoine Venant and François Lareau and Philippe Langlais},

year  = {2025},

date = {2025-09-01},

booktitle = {Proceedings of the 16th International Conference on Computational Semantics (IWCS)},

address = {Düsseldorf, Germany},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{petrov_q-lf_2025,

title = {Q&A-LF: A French Question-Answering Benchmark for Measuring Fine-Grained Lexical Knowledge},

author = {Alexander Petrov and Alessandra Thais Mancas and Viviane Binet and Antoine Venant and François Lareau and Yves Lepage and Philippe Langlais},

year  = {2025},

date = {2025-09-01},

booktitle = {Proceedings of the 15th International Conference on Recent Advances in Natural Language Processing},

address = {Varna, Bulgaria},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{st-gelais_projet_2025,

title = {Le projet Uauitamutau assi. Terminologie de l’environnement en langue innue (innu-aimun)},

author = {Myrriam St-Gelais and Marie-Claude L'Homme and Yvette Mollen},

url = {https://www.uottawa.ca/research-innovation/sites/g/files/bhrskd326/files/2025-06/Programme_Colloque%202025%204.pdf},

year  = {2025},

date = {2025-06-01},

address = {University of Ottawa, Ottawa},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{lhomme_frame_2025,

title = {Frame Semantics for Domain-specific Resources: why it is a good idea and some challenges},

author = {Marie-Claude L'Homme},

url = {The 6th International FrameNet Workshop 2025 (IFNW 2025)},

year  = {2025},

date = {2025-03-01},

address = {Keio University, Yokohama, Japan},

abstract = {Developers of domain-specific resources have referred to Frame Semantics (Fillmore 1982; Fillmore and Baker 2010) and FrameNet (Ruppenhoffer et al. 2016) since the 2000s and have done so in a variety of domains including but not limited to football (Schmidt 2009), adventure tourism (Durán-Muñoz and L’Homme 2020) and medicine (Castaño and Verdaguer 2024). The theoretical and methodological principles on which framenets are based have proven useful for capturing the linguistic properties of terms and connect these properties to the conceptual background of specialized domains. They also help better highlight how knowledge is conceptualized from the perspective of specialized areas, conceptualizations which might differ from the way everyday situations are considered (see Fillmore 1982 on INNOCENT and GUILTY). That said, adaptations are necessary when applying these principles to specialized knowledge which can depend on the field itself or the specific objectives for which resources are developed. 

In this talk, I will first briefly describe the kind of data that domain-specific resource developers typically 

deal with, i.e. single-word and multi-word terms, as well as collocations, and emphasize their efforts to connect this linguistic material to the knowledge structure of specialized domains. I will also say a few words about the methods used most often by developers and some of their limitations. I will then explain why Frame Semantics can help address some of these limitations and show how domain-specific framenets differ from “general-language” ones in terms of lexical content, frames and relations between frames. I will also mention some challenges that domain-specific framenet developers are confronted with, such as heavy reliance on corpus data due to the fact that developers are seldom domain experts, small-size corpora (at least when compared with corpora used by lexicographers) and thus reduced access to tools that learn from large amounts of data. I will focus on the application of 

Frame Semantics and the FrameNet methodology to the development of multilingual resources in the 

field of the environment. The project has been under way since around 2010 (Framed DiCoEnviro 2025; 

L’Homme et al. 2020) and led to the description of terms and frames in English, French, Chinese, Italian, 

Portuguese, and Spanish. 

I will then focus on a new area of research for domain-specific framenet developers, i.e. a method for capturing specialized-frame-evoking multiword units that have been labeled as multiword terms and/or 

collocations (e.g. coastal ecosystem, vulnerable habitat, electric vehicle). Although often compositional, these multiword units have often been listed separately in domain-specific resources. The method I suggest relies on the definition of patterns (loosely based on construction grammar: Hoffmann 2022; Hoffmann and Trousdale 2013) and provides for a generalized account of multiword units in resources. I will conclude by mentioning a few areas that require further investigation but that could still rely on Frame Semantics: capturing differences between text genres in specialized domains (e.g. popularization vs. scientific), a better understanding of domain-specific conceptualizations against “everyday” ones, 

connecting domain-specific framenets to “general” ones, etc.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{petrov_alf_2025-1,

title = {ALF: Un jeu de données d'analogies françaises à grain fin pour l'évaluation de la connaissance lexicale des grands modèles de langue},

author = {Alexander Petrov and Antoine Venant and François Lareau and Yves Lepage and Philippe Langlais},

year  = {2025},

date = {2025-01-01},

booktitle = {Actes de CORIA-TALN-RJCRI-RECITAL 2025.},

volume = {1},

pages = {22–49},

publisher = {ATALA},

address = {Marseille, France},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{marshman_translation_2024,

title = {Translation programs in the age of genAI: Pivotal change, or plus ça change?},

author = {Elizabeth Marshman},

year  = {2024},

date = {2024-06-01},

address = {Montreal},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{gazeau_flexible_2024,

title = {Flexible Lexicalization in Rule-based Text Realization},

author = {Avril Gazeau and François Lareau},

editor = {Nicoletta Calzolari and Min-Yen Kan and Veronique Hoste and Alessandro Lenci and Sakriani Sakti and Nianwen Xue},

url = {https://aclanthology.org/2024.lrec-main.649},

year  = {2024},

date = {2024-05-01},

urldate = {2024-05-31},

booktitle = {Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)},

pages = {7366–7376},

publisher = {ELRA and ICCL},

address = {Torino, Italia},

abstract = {GenDR is a text realizer that takes as input a graph-based semantic representation and outputs the corresponding syntactic dependency trees. One of the tasks in this transduction is lexicalization, i.e., choosing the right lexical units to express a given semanteme. To do so, GenDR uses a semantic dictionary that maps semantemes to corresponding lexical units in a given language. This study aims to develop a flexible lexicalization module to automatically build a rich semantic dictionary for French. To achieve this, we tried two methods. The first one consisted in extracting information from the French Lexical Network, a large-scale French lexical resource, and adapting it to GenDR. The second one was to test a contextual neural language model's ability to generate potential additional lexicalizations. The first method significantly broadened the coverage of GenDR, while the additional lexicalizations produced by the language model turned out to be of limited use, which brings us to the conclusion that it is not suited to perform the task we've asked from it.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liu_assessing_2024,

title = {Assessing BERT's sensitivity to idiomaticity},

author = {Li Liu and François Lareau},

editor = {Archna Bhatia and Gosse Bouma and A. Seza Dogruoz and Kilian Evang and Marcos Garcia and Voula Giouli and Lifeng Han and Joakim Nivre and Alexandre Rademaker},

url = {https://aclanthology.org/2024.mwe-1.4},

year  = {2024},

date = {2024-05-01},

urldate = {2024-05-31},

booktitle = {Proceedings of the Joint Workshop on Multiword Expressions and Universal Dependencies (MWE-UD) @ LREC-COLING 2024},

pages = {14–23},

publisher = {ELRA and ICCL},

address = {Torino, Italia},

abstract = {BERT-like language models have been demonstrated to capture the idiomatic meaning of multiword expressions. Linguists have also shown that idioms have varying degrees of idiomaticity. In this paper, we assess CamemBERT's sensitivity to the degree of idiomaticity within idioms, as well as the dependency of this sensitivity on part of speech and idiom length. We used a demasking task on tokens from 3127 idioms and 22551 tokens corresponding to simple lexemes taken from the French Lexical Network (LN-fr), and observed that CamemBERT performs distinctly on tokens embedded within idioms compared to simple ones. When demasking tokens within idioms, the model is not proficient in discerning their level of idiomaticity. Moreover, regardless of idiomaticity, CamemBERT excels at handling function words. The length of idioms also impacts CamemBERT's performance to a certain extent. The last two observations partly explain the difference between the model's performance on idioms versus simple lexemes. We conclude that the model treats idioms differently from simple lexemes, but that it does not capture the difference in compositionality between subclasses of idioms.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{alfetlawi_can_2024,

title = {Can generative AI help users overcome classic challenges for MT? A pilot study},

author = {Anwar Ghanim Alfetlawi and Elizabeth Marshman},

year  = {2024},

date = {2024-02-01},

address = {Doha, Qatar},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{joan_casademont_ensenanza_2024,

title = {Enseñanza de lenguas adicionales: cómo la investigación científica basada en la validez ecológica puede contribuir a la emancipación digital del profesorado},

author = {Anna Joan Casademont and Nancy Gagné},

year  = {2024},

date = {2024-01-01},

pages = {1821–1825},

publisher = {EDUTEC24},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{michot_carre_2024,

title = {Un carré, est-ce toujours un carré? : pistes pour travailler le lexique mathématique au primaire.},

author = {Sandrine Michot and Ophélie Tremblay},

year  = {2024},

date = {2024-01-01},

booktitle = {Actes du colloque du GDM 2023},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{san_martin_what_2024,

title = {What Generative Artificial Intelligence Means for Terminological Definitions},

author = {Antonio San Martín},

url = {https://ceur-ws.org/Vol-3703/paper1.pdf},

year  = {2024},

date = {2024-01-01},

booktitle = {Proceedings of the 3rd International Conference on Multilingual Digital Terminology Today (MDTT 2024)},

publisher = {CEUR-WS},

address = {Granada},

abstract = {This paper examines the impact of Generative Artificial Intelligence (GenAI) on the creation and consumption of terminological definitions. GenAI tools like ChatGPT present a mix of benefits and drawbacks compared to traditional terminological resources. ChatGPT excels in providing context-specific meanings in an interactive and customized fashion but faces challenges with accuracy. Terminological definitions in recognized resources will likely survive because of their reliability. From the point of view of the terminologist, tools like ChatGPT enable AI-assisted terminography, including post-editing terminography, as an approach blending AI efficiency with human expertise for faster definition creation.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{escouflaire_subjectivite_2024,

title = {La subjectivité dans le journalisme québécois et belge : Transfert de connaissances inter-médias et inter-cultures},

author = {Louis Escouflaire and Antoine Venant and Antoine Descampes and Cédric Fayron},

year  = {2024},

date = {2024-01-01},

booktitle = {JADT 2024: 17th International Conference on Statistical Analysis of Textual Data},

address = {Brussels},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{rigouts_terryn_exploratory_2024,

title = {Exploratory Study on the Impact of English Bias of Generative Large Language Models in Dutch and French},

author = {Ayla Rigouts Terryn and Miryam Lhoneux},

year  = {2024},

date = {2024-01-01},

booktitle = {Proceedings of The 4th Workshop on Human Evaluation of NLP Systems (HumEval’24) at LREC-COLING},

pages = {12–27},

publisher = {European Language Resources Association},

address = {Turin, Italy},

abstract = {The most widely used LLMs like GPT4 and Llama 2 are trained on large amounts of data, mostly in English but are still able to deal with non-English languages. This English bias leads to lower performance in other languages, especially low-resource ones. This paper studies the linguistic quality of LLMs in two non-English high-resource languages: Dutch and French, with a focus on the influence of English. We first construct a comparable corpus of text generated by humans versus LLMs (GPT-4, Zephyr, and GEITje) in the news domain. We proceed to annotate linguistic issues in the LLM-generated texts, obtaining high inter-annotator agreement, and analyse these annotated issues. 

 We find a substantial influence of English for all models under all conditions: on average, 16textbackslash% of all annotations of linguistic errors or peculiarities had a clear link to English. Fine-tuning a LLM to a target language (GEITje is fine-tuned on Dutch) reduces the number of linguistic issues and probably also the influence of English. We further find that using a more elaborate prompt leads to linguistically better results than a concise prompt. Finally, increasing the temperature for one of the models leads to lower linguistic quality but does not alter the influence of English.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}