Papers by Silvia Casola
Reason to Rote: Rethinking Memorization in Reasoning (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models readily memorize arbitrary training instances, such as label noise . however, such memorization does not affect generalizable reasoning abilities . |
| Approach: | They investigate how large language models memorize label noise and why it affects generalizability. |
| Outcome: | The proposed model performs well on reasoning tasks even when memorized labels are missing . the proposed model is able to generalize to correctly answer "87+19=106" |
PERSEVAL: A Framework for Perspectivist Classification Evaluation (2025.emnlp-main)
Copied to clipboard
Soda Marem Lo, Silvia Casola, Erhan Sezerer, Valerio Basile, Franco Sansonetti, Antonio Uva, Davide Bernardi
| Challenge: | Perspectivist evaluation practices in NLP remain fragmented and inconsistent . |
| Approach: | They propose a framework that evaluates perspectivist models at the individual annotator level and treats annotators and users as distinct entities, consistent with real-world scenarios. |
| Outcome: | The proposed framework evaluates annotators and users as distinct entities consistent with real-world scenarios. |
I’m sure you’re a real scholar yourself: Exploring Ironic Content Generation by Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Moreover, irony is highly subjective and can depend on various factors, such as social, cultural, or generational aspects. |
| Approach: | They propose to fine-tune two large language models to generate ironic and non-ironic content and analyze their outputs from a linguistic perspective. |
| Outcome: | The proposed models generate ironic and non-ironic responses to a given social media post and analyze their outputs from a linguistic perspective. |
Confidence-based Ensembling of Perspective-aware Models (2023.emnlp-main)
Copied to clipboard
Silvia Casola, Soda Lo, Valerio Basile, Simona Frenda, Alessandra Cignarella, Viviana Patti, Cristina Bosco
| Challenge: | Human label variability has been a topic of research in the field of NLP recently . Exploiting disagreements in annotations has been shown to offer advantages for accurate modelling and fairer evaluation. |
| Approach: | They propose a highly perspectivist model that exploits disagreements in annotations to capture the subjectivity encoded in the annotation process. |
| Outcome: | The proposed model is validated on irony and hate speech detection scenarios in in-domain and cross-domain settings. |