Papers by Tommaso Caselli
Systems’ Agreements and Disagreements in Temporal Processing: An Extensive Error Analysis of the TempEval-3 Task (L18-1)
Copied to clipboard
| Challenge: | Temporal Processing systems are crucial for timelines and storylines . TempEval-3 is the latest evaluation campaign on open-domain TP in english . |
| Approach: | They present a Temporal Processing system that incorporates high level lexical semantic features and uses them to evaluate temporal relation classification. |
| Outcome: | The proposed system achieves the best scores for event detection and temporal relation classification from raw text, but the errors are not as robust as previous systems. |
Dynamic Stance: Modeling Discussions by Labeling the Interactions (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Stance detection is a popular task that has been modeled as a static task, but its limitations are strong topic-dependent. |
| Approach: | They propose to model stance as a dynamic task by focusing on interactions between a message and their replies. |
| Outcome: | The proposed model shows portability across topics and languages. |
WikiBio: a Semantic Resource for the Intersectional Analysis of Biographical Events (2023.acl-long)
Copied to clipboard
Marco Antonio Stranisci, Rossana Damiano, Enrico Mensa, Viviana Patti, Daniele Radicioni, Tommaso Caselli
| Challenge: | Existing corpora and models for biographical event detection are lacking . Detecting biographical events from unstructured data is a useful task to explore and compare bias in representations of individuals. |
| Approach: | They present a corpus annotated for biographical event detection using 20 Wikipedia biographies and 5 existing corpora to train a model. |
| Outcome: | The proposed model detects all mentions of the target-entity in a biography with an F-score of 0.808 and the entity-related events with an 0.859 score. |
Lexical Popularity: Quantifying the Impact of Pre-training for LLM Performance (2026.eacl-long)
Copied to clipboard
| Challenge: | Large Language Models excel in varied tasks, but their mechanisms remain unclear . current LLMs' development has put this assumption in jeopardy, authors say . |
| Approach: | They examine whether LLMs learn generalized linguistic abstraction or rely on surface-level features that match their pre-training data. |
| Outcome: | The proposed model can learn generalized linguistic abstraction or rely on surface-level features that match their pre-training data. |
Language is Scary when Over-Analyzed: Unpacking Implied Misogynistic Reasoning with Argumentation Theory-Driven Prompts (2024.emnlp-main)
Copied to clipboard
| Challenge: | a new study aims to understand the implicit reasoning used to convey misogynistic comments in Italian and English. |
| Approach: | They propose misogyny detection as an Argumentative Reasoning task and use argumentation theory to build large language models to understand the implicit reasoning used to convey misogany in Italian and English. |
| Outcome: | The proposed task is an argumentative reasoning task in Italian and English. |
Simulating Identity, Propagating Bias: Abstraction and Stereotypes in LLM-Generated Text (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Persona-prompting is a growing strategy to personalize outputs, but its impact on how LLMs represent social groups remains underexplored. |
| Approach: | They investigate whether persona-prompting leads to different levels of linguistic abstraction . they compare 11 persona driven responses to those of a generic AI assistant . |
| Outcome: | The proposed method can be used to personalize outputs, but its impact on how LLMs represent social groups remains underexplored. |
How about Time? Probing a Multilingual Language Model for Temporal Relations (2022.coling-1)
Copied to clipboard
| Challenge: | XLM-R is a multilingual language model for temporal relation classification between events in four languages. |
| Approach: | They propose to use a multilingual language model for temporal relation classification between events in four languages to obtain contextualized embeddings. |
| Outcome: | The proposed model outperforms state-of-the-art models in obtaining competitive results against state- of-the art systems, but lacks suitable encoded information to address this task. |
The Causal News Corpus: Annotating Causal Relations in Event Sentences from News (2022.lrec-1)
Copied to clipboard
Fiona Anting Tan, Ali Hürriyetoğlu, Tommaso Caselli, Nelleke Oostdijk, Tadashi Nomoto, Hansi Hettiarachchi, Iqra Ameer, Onur Uca, Farhana Ferdousi Liza, Tiancheng Hu
| Challenge: | Existing annotation guidelines for event causality focus on only explicit relations or clauses. |
| Approach: | They propose an annotation schema for event causality that addresses these concerns . they annotated 3,559 event sentences from protest event news with labels on whether it contains causal relations or not. |
| Outcome: | The proposed annotation schema for event causality addresses these concerns . it performs well with 81.20% F1 score on test set and 83.46% in 5-folds cross-validation . |
The “r” in “woman” stands for rights. Auditing LLMs in Uncovering Social Dynamics in Implicit Misogyny (2025.findings-emnlp)
Copied to clipboard
| Challenge: | a recent study examined misogynistic expressions in English and Italian . a taxonomy of social dynamics is used to identify misogorical expressions . |
| Approach: | They examine misogynistic expressions in English and Italian using a taxonomy of social dynamics . they find that LLMs struggle to follow instructions and reason in all settings . |
| Outcome: | The results show that misogynistic expressions are more often implicit than openly hostile . the authors show that LLMs struggle to follow instructions and reason in all settings . |
The Circumstantial Event Ontology (CEO) and ECB+/CEO: an Ontology and Corpus for Implicit Causal Relations between Events (L18-1)
Copied to clipboard
| Challenge: | a new ontology for calamity events models semantic circumstantial relations between event classes . a circumstancial relation makes clear "why" something happened, without necessarily predicting it. |
| Approach: | They propose a circumstantial event ontology that models semantic circumstancial relations between event classes . they propose ECB+ annotated corpus for circumstantal relations and a meta model . |
| Outcome: | The proposed model captures that the change yielded by one event explains to people the happening of the next event when observed. |
Norm It! Lexical Normalization for Italian and Its Downstream Effects for Dependency Parsing (2020.lrec-1)
Copied to clipboard
| Challenge: | Existing tools for lexical normalization of social media data are designed with canonical texts in mind, and this makes it difficult to process data in multiple languages. |
| Approach: | They propose to create a lexical normalization dataset for Italian and analyze the inter-annotator agreement for this task. |
| Outcome: | The proposed model improves the parsing of social media data in Italian and shows that it can be used to translate non-standard social media content to canonical language. |
Dead or Murdered? Predicting Responsibility Perception in Femicide News Reports (2022.aacl-main)
Copied to clipboard
| Challenge: | linguistic expressions of gender-based violence can conceptualize the same event from different perspectives by emphasizing certain participants over others. |
| Approach: | They conduct a large-scale perception survey of GBV descriptions from italian newspapers and train regression models that predict the salience of GV participants with respect to different dimensions of perceived responsibility. |
| Outcome: | The proposed model shows that salient focus is more predictable than salient blame, and perpetrators’ salience is more predictable than victims’ salient. |
TEXT-CAKE: Challenging Language Models on Local Text Coherence (2025.coling-main)
Copied to clipboard
| Challenge: | a new evaluation benchmark, TEXT-CAKE, is used to evaluate language models for text coherence detection. |
| Approach: | They propose a benchmark to evaluate language models' ability to detect text coherence . they analyze multilingual and monolingual LMs with varying architectures and parameters . |
| Outcome: | The proposed model outperforms other models on the TEXT-CAKE evaluation benchmark. |
SocioFillmore: A Tool for Discovering Perspectives (2022.acl-demo)
Copied to clipboard
| Challenge: | SOCIOFILLMORE is a multilingual tool which helps to bring to the fore the focus or the perspective that a text expresses in depicting an event. |
| Approach: | They propose a multilingual tool which helps to bring to the fore the focus or the perspective that a text expresses in depicting an event. |
| Outcome: | The proposed tool can be used by non-NLP researchers and is based on a large collection of human judgements. |
Fighting the COVID-19 Infodemic: Modeling the Perspective of Journalists, Fact-Checkers, Social Media Platforms, Policy Makers, and the Society (2021.findings-emnlp)
Copied to clipboard
Firoj Alam, Shaden Shaar, Fahim Dalvi, Hassan Sajjad, Alex Nikolov, Hamdy Mubarak, Giovanni Da San Martino, Ahmed Abdelali, Nadir Durrani, Kareem Darwish, Abdulaziz Al-Homaid, Wajdi Zaghouani, Tommaso Caselli, Gijs Danoe, Friso Stolk, Britt Bruntink, Preslav Nakov
| Challenge: | a dataset of 16K manually annotated tweets is used to analyze disinformation . the democratic nature of social media has raised questions about the quality and the factuality of the information that is shared on these platforms. |
| Approach: | They use a dataset of manually annotated tweets to analyze COVID-19 disinformation . they show that tweets contain fake cures, rumors, conspiracy theories and xenophobia . |
| Outcome: | The proposed dataset shows that it is useful in monolingual vs. multilingual settings. |
I Feel Offended, Don’t Be Abusive! Implicit/Explicit Messages in Offensive and Abusive Language (2020.lrec-1)
Copied to clipboard
| Challenge: | Recent literature suggests different approaches to identify abusive language phenomena . however, there is a lack of data sets that take into account the degree of explicitness . |
| Approach: | They propose to use annotation guidelines to distinguish between explicit and implicit abuse in English and apply them to OLID/OffensEval. |
| Outcome: | The proposed tool distinguishes between explicit and implicit abuse in English and takes into account the degree of explicitness. |