CEUR Proceedings are Online!
http://ceur-ws.org/Vol-3287


Schedule – Wednesday, November 30th, 2022

ROOM W3  (it is on the ground floor of the library building "Biblioteca Rizzi", in Via delle Scienze, Udine)

10:00-10:15 CET

10:00-10:15 – Welcome to NL4AI 2022

10:15-10:45 CET

SESSION 1: Multimodal NLP
Chair: Lucia Passaro

10:15-10:30 – Daniele Borghesi, Andrea Amelio Ravelli and Felice Dell’Orletta. “What makes the audience engaged? Engagement prediction exploiting multimodal features” [10]

10:30-10.45 – Fabio De Ponte and Sarah Rauchas. “Grounding Words in Visual Perceptions: Experiments in Spoken Language Acquisition” [6]

10:45-11:15 CET

10:45 – 11.15 – Coffee Break

11:15-12:00 CET

11:15-12:00 – Invited talk: Fabio Petroni – “Improving Wikipedia Verifiability with AI” – Slides

Abstract: Verifiability is a core content policy of Wikipedia: claims that are likely to be challenged need to be backed by citations. There are millions of articles available online and thousands of new articles are released each month. For this reason, finding relevant sources is a difficult task: many claims do not have any references that support them. Furthermore, even existing citations might not support a given claim or become obsolete once the original source is updated or deleted. Hence, maintaining and improving the quality of Wikipedia references is an important challenge and there is a pressing need for better tools to assist humans in this effort. In the talk, I’ll present our recent research showing that the process of improving references can be tackled with the help of AI. Our results indicate that an AI-based system could be used, in tandem with humans, to improve the verifiability of Wikipedia. More generally, we hope that our work can be used to assist fact-checking efforts and increase the general trustworthiness of information online.

12:00-13:00 CET

SESSION 2: Applied NLP
Chair: Debora Nozza

12.00-12.15 – Valerio Basile. “Is EVALITA Done? On the Impact of Prompting on the Italian NLP Evaluation Campaign.” [13]

12.15-12.30 – Elena Bellodi, Alessandro Bertagnon and Marco Gavanelli. “Comparing Emotion and Sentiment Analysis Tools on Italian anti-vaccination for COVID-19 posts” [9]

12.30-12.45 – Alex J. Lucassen, Alberto Testoni and Raffaella Bernardi. “Probability Distributions as a Litmus Test to Inspect NNs Grounding Skills” [11]

12.45-13.00 – Hamza Tarik Sadouk, Faouzi Sebbak, Walid Cherifi and Mohamed Amine Chouarfia. “Source Reliability Estimation for the Verification of the Authenticity of Information: an Evidential Approach” Canceled

13:00-14:00 CET

12:45-14:00 – Lunch Break

14:00-14:40 CET

SESSION 3: Selected Papers
Chair: Marco Polignano

14:00-14:20 – David La Barbera, Kevin Roitero and Stefano Mizzaro. “A Hybrid Human-In-The-Loop Framework for Fact Checking” [4]

14:20-14:40 – Claudiu Daniel Hromei, Danilo Croce and Roberto Basili. “Grounding end-to-end Architectures for Semantic Role Labeling in Human Robot Interaction” [5]

14:40-15:00 CET

14:40-15:00 – Short Break

15:00-15:45 CET

SESSION 4: NLP and Structured Languages
Chair: Debora Nozza

15:00-15:15 – Muhammad Saad Amin, Alessandro Mazzei and Luca Anselma. “Towards Data Augmentation for DRS-to-Text Generation” [14]

15:15-15:30 – Irene Sucameli, Alessandro Bondielli, Lucia Passaro, Edoardo Annunziata, Giulia Lucherini, Andrea Romei and Alessandro Lenci. “MATE, a Meta Layer Between Natural Language and Database” [15]

15:30-15:45 – Patrizio Bellan, Mauro Dragoni, Simone Paolo Ponzetto, Chiara Ghidini and Han van der Aa. “Process Extraction from Natural Language Text: the PET dataset and annotation guidelines” [18]

15:45-16:15 CET

15:45-16:15 – Coffee Break

16:15-17:00 CET

SESSION 5: Challenges in NLP
Chair: Marco Polignano

16:15-16:30 – Michele Dusi, Nicola Arici, Alfonso Emilio Gerevini, Luca Putelli and Ivan Serina. “Graphical Identification of Gender Bias in BERT with a Weakly Supervised Approach” [16]

16:30-16:45 – Tiziano Labruna and Bernardo Magnini Simulating Domain. “Changes in Conversational Agents Through Dialogue Adaptation” [17]

16:45-17:00 – Michele Papucci, Chiara De Nigris, Alessio Miaschi and Felice Dell’Orletta. “Evaluating Text-To-Text Framework for Topic and Style Classification of Italian texts” [8]

17:00-17:45 CET

17:00-17:45 – Invited talk: Giuseppe AttardiLarge Language Models are All You Need Slides

Abstract: LLMs are among the three scientific breakthroughs of Deep Learning applied to NLP in just ten years: word embeddings, attention, and prompt engineering. They have shown surprising effectiveness in all tasks and have also been adopted in other areas, including cross-cutting tasks such as generating images from text descriptions. However, LLMs have also raised perplexities, starting with doubts about the release of GPT-2 to Galactica’s recent retirement. Their capabilities, which seem to grow with their size, have become a subject of study. We will discuss whether discrimination will continue to grow among groups of researchers capable of building them and about possible alternatives.

17:45-18:00 CET

17:45-18:00 – Closing Remarks