Workshop Italian Visual and Language models: challenges and activities

General Information

When: February, 5th 2024
Where: AGO Modena Fabbriche Culturali - Piazzale Sant'Agostino Modena (map)

This second workshop, after the one organized in Rome, focuses on the development of Large Multimodal Models (LMMs) for the Italian language within the framework of the Transversal Project "Vision, Language, and Multimodal Challenges" as part of the project "Future Artificial Intelligence Research" (FAIR). The goal of the event is to discuss in the Italian Community the future challenges on design, development, evaluation of Large Language Models, Large Visual Models and their combination as well different application domains such as Health or Manufacture. During the workshop some results of FAIR research activities will be presented.

Organizers

• Prof. Rita Cucchiara (CNR-UNIMORE)
• Prof. Roberto Navigli (La Sapienza)

Agenda

09:00-09:10 Registration
09:10-09:30 Welcome from the President of the Emilia-Romagna region (Stefano Bonaccini), the Rector of the University of Modena and Reggio Emilia (Carlo Adolfo Porro) and the President of the Fondazione Ago (Donatella Pieri)
09:30-9:45 Presentation of the Project and the Results of the Rome Workshop (R. Navigli)
09:45-11:00 Presentation of challenges/activities of Large Visual and Language Models of FAIR (R. Cucchiara, R. Navigli and all FAIR)
Coffee break
11:30-12:00 Presentation of challenges/activities of Large Visual and Language Models of FAIR (R. Cucchiara, R. Navigli and all FAIR) (continuation)
12:00-12:30 Sanzo Bassini (Director of SuperComputing Applications and Innovation Department, CINECA) "The Present and Future of the collaboration between FAIR and CINECA"
12:30-13:00 Invited speech: Massimo Sideri (Corriere della Sera) "AI e il Visconte Cibernetico"
13:00-14:30 Lunch (tortellini guaranteed) and Poster Presentation
14:30-15:00 Invited speech: Jenia Jitsev and Marianna Nezhurina (LAION) "Open foundation models: reproducible science of transferable learning"
15:00-15:30 Presentation of FAIR results on LLM and LMM

• Focus on Visual datasets: Giovanni Maria Farinella
• Focus on Italian Visual and Language prototypes (the MORE model): Lorenzo Baraldi
• Focus on Italian Large Language Models (LLaMAntino): Pierpaolo Basile, Giovanni Semeraro

15:30-17:00 Discussion on the following topics with contributions from the participants
17:00-17:30 Conclusions

Discussion topics

- Challenges on "multimodal adapters" and Foundation models
- Challenges on application domains and datasets (with participation of stakeholders companies)
- Challenges on Evaluation and Trustworthiness Assessment
- Preparation of a FAIR Position Paper on Multimodal Foundation Models
- Next project activities.

Contacts and registration

Registration for in-person participation is closed.. You can still register for on-line attendance by sending an e-mail to one of the addresses below. The URL will be sent to registered participants the day of the workshop.

For information, please contact segreteria.aimagelab@unimore.it or lorenzo.baraldi@unimore.it.

The workshop is supported by the National PhD in "AI for Society", sede di Modena.