D1.8: Final ten-pilots validation report and lessons learned (V1.0) - Archive ouverte HAL
Rapport Année : 2024

D1.8: Final ten-pilots validation report and lessons learned (V1.0)

Enrico Daga
Marco Gurrieri
Simon Holland
  • Fonction : Auteur
  • PersonId : 1111288

Résumé

This deliverable is the final deliverable on Pilots in Workpackage 1 of Polifonia, next to the deliverable D1.10 about the Polifonia Web Portal. WP 1 (Pilots and Web portal) has been also called Polifonia’s validation WP. It contributes to achieve the objectives O3 (Tailor) and O5 (Share & Engage)1. As such it drives the whole development of Polifonia and in particular the monitoring and aligning of the 10 Pilots. This last deliverable builds on previous deliverables in this WP. Among them D1.1 Roadmap and pilot requirements 1st version [1], D1.2 Roadmap and pilot requirements – 2nd version [2], and D1.3 Pilots development - collaborative methodology and tools [3] defined the collaborative strategy of the project, the main knowledge sources and epistemological units (research questions), and how to best monitor and management the overall progress in an otherwise distributed knowledge exploration process. In this process each of the Pilots can be envisioned as a scout send out to explore other dimensions of the actual problem space. But, to achieve a scientific progress in such an interdisciplinary setting as Polifonia represents, the knowledge colected by those scouts needs to be brought together again and integrated. Polifonia did this in an iterative way, and dedicated a specific governance board - the Technical Board - to this task. The individual Pilots each on its own way also operated on fronts of research, methods, and data curation. For instance for the more methodologically operating Pilots (scope Studying) the work crucially includes cutting-edge experiments evaluating Large Language Models (LLM) in the context of various scenarios relevant to the stakeholders. The latter belongs to the unforeseen new challenges during the journey of the project. Another central challenge concerns how the consortium has been experimenting on how symbolic knowledge (Knowledge Graphs) can interplay with predictive, generative models. To both acknowledge the Pilot specific achievements and the contribution to the project-wide goals, this deliverable devotes one section to each Pilot, but uses the same kind of template (subsection headings) to report about the Pilots. In their report the focus lies on main advances and the validation of results. Moreover, each Pilot description details on which other (more elaborate Deliverables) this work is based. In a last section, this deliverable includes discussion of the reusability of Polifonia’s research and outputs. We have gathered data on the reusability of different components of the Polifonia ecosystem via a self-reporting survey, and we report on this in Chapter 12. Further, we revisit a table of anticipated reusability of components, originally presented in the Polifonia Grant Agreement, and map it against our outcomes.
Fichier principal
Vignette du fichier
D1_8___Final_ten_pilots_validation-V1.0-RC3.pdf (19.83 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04582760 , version 1 (22-05-2024)

Identifiants

  • HAL Id : hal-04582760 , version 1

Citer

Nicholas Canny, Jason Carvalho, Alba Catalina Morales Tirado, Enrico Daga, Jacopo de Berardinis, et al.. D1.8: Final ten-pilots validation report and lessons learned (V1.0). Open University; King's College London; IReMus; CNAM; NUIG; KNAW; MiC; Università di Bologna. 2024. ⟨hal-04582760⟩
112 Consultations
35 Téléchargements

Partager

More