August, 2022.- Cinthia Sánchez, a PhD candidate in Computing at the Department of Computer Science at the University of Chile and an IMFD student, is investigating the problem of data scarcity, in order to create new multilingual and multidomain models. to classify messages related to crises, which allow their generalization to different critical tasks in crisis management.
Her research A Multilingual and Multi-Domain Approach for Crisis Classification in Social Media is part of the Crisis Informatics area, where works and tools are found that support emergency management based on information published on social networks.
The focus of her work together with Bárbara Poblete, DCC U. Chile academic and IMFD researcher, and Andrés Abeliuk, DCC U. Chile academic, is the scarcity of data, and seeks to improve the scope of automated analysis of social networks during emergencies. , through natural language processing, knowledge transfer learning techniques and classification algorithms, both traditional and based on neural networks.
“Crisis Informatics presents interesting challenges to which to contribute, such as the detection of relevant and actionable information in multiple languages, and the creation of models and tools that undoubtedly have a direct impact on society”, explains Sánchez.