Papers by Olga Seminck
FReND: A French Resource of Negation Data (2024.lrec-main)
Copied to clipboard
| Challenge: | Negation data are limited by the language models of the BERT-generation, which are still underperforming on tasks and benchmarks featuring negation. |
| Approach: | FReND is a freely available corpus of French language in which negations are hand-annotated by their cues and scopes. |
| Outcome: | FReND is the largest dataset available for french negation analysis . it is a valuable resource for linguistic research and as training data for AI tasks such as negation detection. |
Subject Verb Agreement Error Patterns in Meaningless Sentences: Humans vs. BERT (2022.coling-1)
Copied to clipboard
| Challenge: | Existing research shows that humans are prone to making agreement errors with specific constructions. |
| Approach: | They compare the performance of BERT-base and that of humans using crowdsourcing . they find that meaningfulness is stronger for BERT than for humans . |
| Outcome: | The proposed model performs better than humans on a crowdsourcing experiment . |
A Gold Anaphora Annotation Layer on an Eye Movement Corpus (L18-1)
Copied to clipboard
| Challenge: | Anaphora resolution is a complex process in which multiple linguistic factors play a role. |
| Approach: | They used annotated anaphorical pronouns from newspaper articles read by humans to model reading time of pronounes. |
| Outcome: | The proposed resource allows to study human anaphora resolution on natural data. |