Agent-based Modeling and Synthetic Populations for Social Research
Text as Data: The Potential of Large Language Models for Social Research
- Published on
- Last updated
This event has ended.
Registrazione video di Text as Data: The Potential of Large Language Models for Social Research
Aggiornamento:
- Guarda la registrazione del corso sul canale YouTube del CNR DSU
- Ottieni il materiale condiviso durante il corso
Il Corso
Questo corso introduce i Large Language Models (LLM) come strumenti emergenti per la ricerca nelle scienze sociali, mettendo in evidenza la loro capacità di integrare dati testuali e strutturati all’interno di quadri analitici.
Il corso sarà tenuto in lingua inglese
This course introduces Large Language Models (LLMs) as emerging tools for social science research, emphasizing their capacity to integrate textual and structured data within analytical frameworks. Building on transformer architectures, LLMs can represent complex linguistic information through embeddings – dense numerical vectors that capture semantic, contextual, and relational meaning from text. These representations allow textual data from diverse sources – such as policy documents, survey open-ended responses, scientific publications, administrative records or social media content – to be combined with traditional quantitative variables in empirical models. The course discusses how this integration can enhance prediction, interpretation, and causal inference in social research, offering a conceptual and practical overview of how foundation models reshape the “text-as-data” paradigm.
Relatrice
Loreta Isaraj (CNR-IRCrES)
Come partecipare
Il corso è online e aperto al pubblico previa registrazione attraverso questo modulo.
Comitato organizzatore
Tommaso Ismaelli, Andrea Giovanni Nuzzolese.
