Technology
Multi-Agent Model
This research project will have to cope with a major obstacle: starting from a potentially variable stream of raw data, I-Muse App will have to be able to infer both individual visitor tastes and future stream of visitors for each museum participating. We will be thus dealing with data that that summarize complex phenomena which are strongly influenced by human actions. In this context, we are developing our system following a multi-agent paradigm: different structural entities (the agents) perform AI-powered tasks managed by a special entity (the manager), and communicating through a shared database. The result is a flexible, scalable, robust, and efficient ecosystem. Flexible, to allow future researchers to insert additional agents without creating conflict with the existing ones; scalable, to allow more museums to enter in the collaboration and provide more information regarding their artworks and visits; robust, to handle data that may be incomplete or partially incorrect; efficient, to coordinate the access to shared computational resources, such as databases and processing units.
Art Embeddings
The way we access cultural content is progressively shaped by the use of Artificial Intelligence technologies that can create personalized experiences for users. These systems generally relate artworks expressed on the same medium – e.g. paintings, music or movies. However, cultural and artistic expressions are developed according to different paradigms, and our natural approach to culture is not necessarily related to its support. Works of art are rarely considered as stand-alone entities, they are a product of the historical period in which they were conceived. Hence, we are exploring the capabilities of AI technologies in creating embeddings of heterogeneous art pieces of Turin’s museums, displaying links that may transcend the individual items’ peculiarities. The aim of this work is to innovate the experience of visitors in museums, reinforcing the idea of computational intelligence as an important tool for cultural dissemination.
Predictions
One of the main objectives of this project is to provide both museums and visitors with information on the amount of people that are expected to visit at future dates. To do so we decided to implement different methodologies and statistical data generating processes. A wide variety of machine learning models has been considered to achieve this task, such as regularized regressions, support vector machines, ARIMA and SARIMAX models, as well as ensemble methods – random forest and gradient boosting regressors. The models are to be consequently integrated in the multi-agent model ecosystem – as explained in the above section, where each model/agent will output predictions for future dates and the manager will then collect the predictions in order to provide the most accurate forecast with the available information.
Academic research output
Coming soon…
Who?
The origins of Fondazione Compagnia di San Paolo (FCSP), one of Europe’s largest private foundations of banking origin, go back to a confraternity founded in 1563. The Foundation’s structure has been transformed several times over the centuries, to meet the needs of citizens. Throughout its evolution, it has upheld its philanthropic calling and continued to provide help and support for the civil, cultural and economic development of the community in which it operates. The decision to adopt these three goals stems from FCSP’s desire to align itself with the international shared framework of the United Nations, which established the 2030 Agenda in 2015, asking individual countries to adopt the 17 Sustainable Development Goals (SDGs) and the associated 169 targets, as well as the whole philosophy behind them.
Why?
The origins of Fondazione Compagnia di San Paolo (FCSP), one of Europe’s largest private foundations of banking origin, go back to a confraternity founded in 1563. The Foundation’s structure has been transformed several times over the centuries, to meet the needs of citizens. Throughout its evolution, it has upheld its philanthropic calling and continued to provide help and support for the civil, cultural and economic development of the community in which it operates. The decision to adopt these three goals stems from FCSP’s desire to align itself with the international shared framework of the United Nations, which established the 2030 Agenda in 2015, asking individual countries to adopt the 17 Sustainable Development Goals (SDGs) and the associated 169 targets, as well as the whole philosophy behind them.The growing popularity of Artificial Intelligence and the Internet of Things represent an invaluable opportunity for both individuals and society as a whole. The widespread use of these disciplines and the related methods is shaping new approaches to working habits, education, management of economic and cultural activities, as well as the social context in which they are embedded. In the wake of these changes, The Compagnia di San Paolo Foundation published the tender initiative “Artificial Intelligence, Humans and Society” in January 2020, with the aim to support innovative research projects in the field of Artificial Intelligence providing for social and economic benefits to local communities.