Descrizione del progetto
Uno studio scientifico dell’evoluzione letteraria che integra intelligenza artificiale, modellizzazione statistica e megadati
I megadati e la digitalizzazione offrono possibilità per enormi informazioni approfondite in campi che spaziano ben oltre la scienza e la tecnologia. Attualmente, i lettori usufruiscono della letteratura su dispositivi elettronici e i loro pensieri relativi a ciò che hanno letto compaiono su piattaforme digitali che comprendono siti di recensioni e social media. Il progetto GOLEM, finanziato dal CER, beneficerà di dati senza precedenti, tra cui la lettura di storie e gli effetti di queste ultime sui lettori così come compresi dai commenti sui mezzi digitali e i social media. Avvalendosi dell’apprendimento automatico e degli strumenti di modellizzazione statistici, nonché confrontando le relazioni tra le storie in cinque diverse lingue, il team creerà modelli accurati che getteranno luce su come la narrativa si adatti alle circostanze sociali e alle culture.
Obiettivo
The “Graphs and Ontologies for Literary Evolution Models” (GOLEM) project will create statistically robust models explaining how fiction evolves, based on the analysis of millions of stories and the effects they have on readers. This is the first time in history that this kind of data is available on such a large scale, thanks to the fact that readers all over the world use digital and social media to share fictional stories and to comment on them, e.g. on fanfiction websites or on publishing platforms like Wattpad. GOLEM will use computational literary studies and cultural evolution theory to create accurate models of how the (formal and content-related) cultural traits found in fiction spread and combine. The basis of this evolutionary analysis of fiction will be a knowledge graph database – an infrastructure of interlinked data about stories and reader response – which will be used to test hypotheses related to the accumulation of cultural traits in stories and their effectiveness in achieving cognitive and emotional effects on readers. State-of-the-art machine learning algorithms and advanced statistical modelling tools will be employed to create a major breakthrough in computational literary studies, possibly also contributing to the revision of cultural evolution theories. By focusing on the relations between stories in five different languages, collected from countries in all continents, GOLEM will provide an unprecedented insight into how storytelling, one of the most ancient cultural systems, evolves. Literary history and criticism have offered refined accounts of how fiction works, mostly relying on case studies of limited extent. It is now time to provide robust statistical evidence of the anthropological function of fiction and of how it adapts to different circumstances and cultures, empowering readers to cope with their cultural or societal contexts.
Campo scientifico
Parole chiave
Programma(i)
- HORIZON.1.1 - European Research Council (ERC) Main Programme
Argomento(i)
Meccanismo di finanziamento
ERC - Support for frontier research (ERC)Istituzione ospitante
9712CP Groningen
Paesi Bassi