Table des matières
Keynote Talks and Tutorial
Barbara Plank
Returning the L in NLP: Why Language (Variety) Matters and How to Embrace it in Our ModelsDan Roth
It’s Time for ReasoningJulia Bosque-Gil
Introduction to Linguistic Linked DataDaniel Zeman
Universal Dependencies: Principles and ToolsContributed Papers: Long Papers
Akshay Aggarwal et Chiara Alzetta
Atypical or Underrepresented? A Pilot Study on Small TreebanksGiorgia Albertin, Alessio Miaschi et Dominique Brunato
On the Role of Textual Connectives in Sentence Comprehension: A New Dataset for ItalianLuca Anselma, Dario Ferrero et Alessandro Mazzei
Query in linguaggio naturale per il dominio della dieta mediterranea- 1. Introduzione
- 2. Alcuni lavori recenti LN- - >SQL basati sulla sintassi
- 3. Dominio applicativo: il progetto MAdiMan
- 3.1 La base di dati: Gedeone
- 4. Architettura del prototipo LN- - >SQL
- 4.1 Prima fase: dal Linguaggio Naturale al Calcolo Relazionale
- 4.2 Seconda fase: dal Calcolo Relazionale a SQL
- 5. Valutazione preliminare
- 6. Conclusioni
Lavinia Aparaschivei, Lionel Nicolas et Alberto Barrón-Cedeño
Are Crescia and Piadina the Same? Towards Identifying Synonymy or Non-Synonymy between Italian Words to Enable Crowdsourcing from Language LearnersChiara Barattieri di San Pietro, Marco Marelli et Carlo Reverberi
Moving from Human Ratings to Word Vectors to Classify People with Focal Dementias: Are We There Yet?Silvia Casola et Alberto Lavelli
WITS: Wikipedia for Italian Text SummarizationFabio Celli, Mirko Lai, Armend Duzha et al.
Policycorpus XL: An Italian Corpus for the detection of Hate Speech Against PoliticsAldo Cerulli, Dominique Brunato et Felice Dell’Orletta
Quale testo è scritto meglio? A Study on Italian Native Speakers’ Perception of Writing QualityMauro Cettolo, Roldano Cattoni et Marco Turchi
On the Development of Customized Neural Machine Translation ModelsCristiano Chesi
A Common Derivation for Parsing and Generation with Expectation-Based Minimalist Grammars (e-MGs)Alessandra Teresa Cignarella, Mirko Lai, Andrea Marra et al.
“La ministro è incinta”: A Twitter Account of Women’s Job Titles in ItalianDanilo Croce, Lucia C. Passaro, Alessandro Lenci et al.
GQA-it: Italian Question Answering on Image Scene GraphsAnna Dall’Acqua et Fabio Tamburini
Implementing a Pragmatically Adequate Chatbot in DialogFlow CX- 1. Introduction
- 2. The Natural Conversation Framework as a Theoretically Funded Approach
- 3. Dialogflow CX
- 4. Conversational Architecture and Pattern Selection
- 4.1 Default Start Flow
- 4.2 Registration (Registrazione)
- 4.3 App Download (Scaricare App)
- 4.4 Discounts (Sconti)
- 5. Conclusions and Future Directions of the Research
Maria De Martino et Annalisa Colella
La produzione di testi semplificati di notiziari televisivi italiani destinati a persone con disturbi cognitivi acquisiti: un’integrazione tra metodi psicolinguistici e analisi automatiche- 1. Introduzione
- 2. Metodo
- 2.1 Selezione dei testi originali delle notizie
- 2.2 Produzione di versioni semplificate delle notizie
- Parametri manipolati per produrre la versione S1
- Parametri manipolati per produrre la versione S2
- 2.3 Analisi delle notizie selezionate
- 2.4 Verifica sperimentale-pilota basata sulla prestazione di parlanti sani
- Partecipanti
- Stimoli
- E1: Procedura
- E1: Risultati
- E2: Procedura
- E2: Risultati
- 3. Conclusioni
Martina Di Bratto, Maria Di Maro, Antonio Origlia et al.
Dialogue Analysis with Graph Databases: Characterising Domain Items Usage for Movie RecommendationsMartina Di Bratto, Riccardo Orrico, Ancuta Budeanu et al.
Do You Have any Recommendation? An Annotation System for the Seekers’ Strategies in Recommendation DialoguesLuca Di Liello, Olga Uryupina et Alessandro Moschitti
Language Transfer for Identifying Diagnostic Paragraphs in Clinical Notes- 1. Introduction
- 2. State of the Art
- 3. Data for Identifying Diagnoses and Procedures Segments
- 3.1 English Data: MIMIC-III
- 3.2 Italian Data: Exprivia Datasets
- 4. Transformer-Based Architectures for Diagnosis and Procedure Extraction
- Tokenization
- Pre-training and fine-tuning
- 5. Language Transfer for Diagnosis Identification
- 6. Experiments
- 6.1 Setup
- Data processing
- Transformer Models
- Evaluation metrics
- 6.2 Results
- Monolingual results
- Cross-lingual experiments
- 7. Conclusion
Irene Dini, Felice Dell’Orletta, Fabio Ferri et al.
Trattamento automatico della lingua a supporto dell’editoria: primi esperimenti con il Devoto-Oli Junior- 1. Introduzione
- 2. Il prodotto Dizionario
- 3. La nuova edizione di un dizionario
- 3.1 La complessità del linguaggio
- 3.2 Il lemmario
- 4. Il ruolo del TAL nella revisione del DJ
- 4.1 Analisi delle definizioni
- 4.2 Revisione del lemmario
- 5. Elaborazioni: alcuni esempi
- 5.1 Complessità lessicale
- 5.2 Complessità sintattica
- 6. Conclusioni
Samuel Fabrizi, Maria Iacono, Andrea Tesei et al.
A First Step Towards Automatic Consolidation of Legal Acts: Reliable Classification of Textual ModificationsKomal Florio, Valerio Basile et Viviana Patti
Hate Speech and Topic Shift in the Covid-19 Public Discourse on Social Media in ItalySimona Frenda, Alessandra Teresa Cignarella, Marco Antonio Stranisci et al.
Recognizing Hate with NLP: The Teaching Experience of the #DeactivHate Lab in Italian High SchoolsEmiliano Giovannetti, Davide Albanesi, Andrea Bellandi et al.
The Role of a Computational Lexicon for Query Expansion in Full-Text SearchFrancesca Grasso et Luigi Di Caro
A Methodology for Large-Scale, Disambiguated and Unbiased Lexical Knowledge Acquisition Based on Multilingual Word Alignment- 1. Introduction
- 2. Background and Related Work
- 2.1 Bias Types
- Linguistic bias
- Lexicographic bias
- 2.2 Related Work
- 3. The Multilingual Word Alignment
- 3.1 Languages Involved
- 3.2 Method
- 4. Implementation
- 4.1 Context for Multilingual Alignment
- 4.2 Multilingual Alignment
- 4.3 Evaluation
- 5. Conclusions and Future Work
Francesco Grotto, Rachele Sprugnoli, Margherita Fantoli et al.
The Annotation of Liber Abbaci, a Domain-Specific Latin ResourceLennert Jansen, Arabella Sinclair, Margot J. van der Goot et al.
Detecting Age-Related Linguistic Patterns in Dialogue: Toward Adaptive Conversational SystemsTiziano Labruna et Bernardo Magnini
From Cambridge to Pisa: A Journey into Cross-Lingual Dialogue Domain Adaptation for Conversational Agents- 1. Introduction
- 2. Dialogue Domain Adaptation
- 2.1 Slot-Value Substitution
- 3. Method
- 3.1 Delexicalization
- 3.2 Translation
- 3.3 Slot-Value Substitution
- 4. Experimental Setting
- Pisa KB-target
- Automatic translation
- Datasets
- Dialogue State Tracking Model
- 5. Results
- 6. Discussion
- 6.1 Overlaps Between Datasets
- 7. Conclusion
Agnese Lombardi et Alessandro Lenci
Agentività e telicità in GilBERTo: implicazioni cognitiveSamuel Louvan, Silvia Casola et Bernardo Magnini
Investigating Continued Pretraining for Zero-Shot Cross-Lingual Spoken Language Understanding- 1. Introduction
- 2. Continued Pretraining in Zero-Shot SLU
- Intermediate Data for Continued Pretraining
- 3. Experimental Setup
- 3.1 Models
- Continued Pretraining
- Fine-Tuning
- Implementation & Model Evaluation metric
- 3.2 Dataset
- SF and IC Dataset
- Continued Pretraining Dataset
- 4. Results
- 5. Analysis and Discussion
- 5.1 Performance Variation when Applying DAPT
- 5.2 Domain Relevance for DAPTEn
- 6. Related Work
- Zero-Shot Cross-Lingual SLU
- Continued Pre-training
- 7. Conclusion
Francesco Mambrini, Eleonora Litta, Marco Passarotti et al.
Linking the Lewis & Short Dictionary to the LiLa Knowledge Base of Interoperable Linguistic Resources for LatinFrancesca Masini, Simone Mattiola et Stefano Dei Rossi
The ListTyp DatabaseAlessio Miaschi, Chiara Alzetta, Dominique Brunato et al.
Probing Tasks Under PressureGosse Minnema, Sara Gemelli, Chiara Zanchi et al.
Frame Semantics for Social NLP in Italian: Analyzing Responsibility Framing in Femicide News Reports- 1. Introduction
- 2. Semantic Frames for Events in Society
- 3. Frame Semantic Parsing for Italian
- 3.1 LOME experiments
- Strategies
- Evaluation
- Implementation
- Results
- 3.2 Evaluating Real-World Performance
- Femicide annotation
- Generalization
- 4. Frame-Based Analysis of Femicide News
- Dataset
- Analysis
- 5. Conclusions
Tolúlọpẹ Ògúnrẹ̀mí, Nazanin Sabri, Valerio Basile et al.
Leveraging Bias in Pre-Trained Word Embeddings for Unsupervised Microaggression DetectionTeresa Paccosi et Alessio Palmero Aprosio
REDIT: A Tool and Dataset for Extraction of Personal Data in Documents of the Public Administration DomainTeresa Paccosi, Alessio Palmero Aprosio et Sara Tonelli
It Is MarkIT That Is New: An Italian Treebank of Marked ConstructionsKaterina Papantoniou, Panagiotis Papadakos, Giorgos Flouris et al.
Linguistic Cues of Deception in a Multilingual April Fools’ Day ContextSara Papi, Matteo Negri et Marco Turchi
Visualization: The Missing Factor in Simultaneous Speech Translation- 1. Introduction
- 2. Background
- 3. Architectural Challenges
- Encoding Strategy
- Decoding Strategy
- Offline or Online training?
- Another point of view: re-translation
- Decision Policy
- Retrospective
- 4. Evaluation Challenges
- Latency Metrics for SimulMT
- Latency Metrics for SimulST
- 5. The Missing Factor: Visualization
- Visualization modalities
- 6. Conclusions and Future Directions
Andrea Amelio Ravelli, Andrea Cimino et Felice Dell’Orletta
Audience Engagement Prediction in Guided Tours through Multimodal FeaturesGuido Rocchietti, Flavia Achena, Giuseppe Marziano et al.
FANCY: A Diagnostic Data-Set for NLI ModelsVeronica Juliana Schmalz et Alessio Brutti
Automatic Assessment of English CEFR Levels Using BERT EmbeddingsVeronica Juliana Schmalz, Jennifer-Carmen Frey et Egon W. Stemle
Introducing a Gold Standard Corpus from Young Multilinguals for the Evaluation of Automatic UD-PoS Taggers for ItalianLucia Siciliani, Pierluigi Cassotti, Pierpaolo Basile et al.
Extracting Relations from Italian Wikipedia Using Self-TrainingRachele Sprugnoli, Francesco Mambrini, Marco Passarotti et al.
Sentiment Analysis of Latin Poetry: First Experiments on the Odes of HoraceSilvia Terragni et Elisabetta Fersini
OCTIS 2.0: Optimizing and Comparing Topic Models in Italian is Even Simpler!- 1. Introduction
- Contributions
- 2. OCTIS: Optimizing and Comparing Topic Models is Simple!
- 2.1 OCTIS 1.0
- Pre-processing and Datasets
- Single-Objective Hyper-parameter Optimization
- 2.2 OCTIS 2.0
- New dataset resources for the Italian language
- From Single-objective to Multi-objective Hyper-parameter Bayesian Optimization
- 3. Experimental Setting
- 3.1 Evaluation Metrics
- IRBO
- NPMI
- 3.2 Topic Models and Hyper-Parameter Setting
- 4. Results
- 4.1 Quantitative Results
- 4.3 Qualitative Results
- 5. Conclusion
Rocco Tripodi
How Contextualized Word Embeddings Represent Word SensesViviana Ventura et Elisabetta Jezek
ConteCorpus: An Analysis of People Response to Institutional Communications During the Pandemic- 1. Introduction
- 2. State of the Art
- 3. ConteCorpus Construction
- 3.1 Data Collection
- 3.2 Processing with the Neural Pipeline Stanza
- 4. End-to-End Evaluation
- 4.1 Construction of the Gold Standard
- 4.2 Evaluation with CoNLL 2018 UD Shared Task Official Evaluation Script
- 5. Topic Modeling
- 5.1 Methodology
- 5.2 Results
- 6. Concluding Observations
Vorakit Vorakitphan, Elena Cabrio et Serena Villata
PROTECT – A Pipeline for Propaganda Detection and Classification- 1. Introduction
- 2. Related Work
- 3. Propaganda Detection and Classification
- 3.1 Datasets
- 3.5 PROTECT Architecture
- Step 1: Propaganda Snippet Detection
- Step 2: Propaganda Technique Classification
- 4. PROTECT Functionalities
- 4.1 Service 1: Propaganda Techniques Classification
- 4.2 Service 2: Propaganda Word Clouds
- 5. Conclusions
Andrea Zugarini et Marco Ernandes
A Multi-Strategy Approach to Crossword Clue Answer Retrieval and RankingContributed Papers: Short Papers
Pierluigi Cassotti, Andrea Iovine, Pierpaolo Basile et al.
Emerging Trends in Gender-Specific Occupational Titles in Italian NewspapersElisabetta Fersini, Luca Rosato, Antonio Candelieri et al.
Deep Learning Representations in Automatic Misogyny Identification: What Do We Gain and What Do We Miss?Maria Iacono, Laura Rossi, Paolo Dangelo et al.
An Obligations Extraction System for Heterogeneous Legal Documents: Building and Evaluating Data and ModelRoberta Bianca Luzietti, Niccolò Pretto, Frédéric Kaplan et al.
FONTI 4.0: Evaluating Speech-to-Text Automatic Transcription of Digitized Historical Oral SourcesSeyed Mahed Mousavi, Roberto Negro et Giuseppe Riccardi
An Unsupervised Approach to Extract Life-Events from Personal Narratives in the Mental Health DomainGennaro Nolano, Mohammad Fazleh Elahi, Maria Pia di Buono et al.
An Italian Question Answering System Based on Grammars Automatically Generated from Ontology LexicaDaniele Puccinelli, Silvia Demartini et Pier Luigi Ferrari
Tackling Italian University Assessment Tests with Transformer-Based Language ModelsLeonardo Ranaldi, Michele Mastromattei, Dario Onorati et al.
KERMIT for Sentiment Analysis in Italian Healthcare ReviewsEmma Romani, Valerio Gattero et Elisabetta Jezek
T-PAS Scraper: an Application for Linguistic Data Extraction and Analysis