Research Grants | CRiSS-LAB

Collective Memory Decay in Science

Wed, 25 Feb 2026 00:00:00 +0000

This ANID-FONDECYT Regular 2026 project studies how science remembers and forgets retracted research. The project asks why invalidated studies can continue to be cited and influence scientific agendas after formal correction.

The research combines large-scale bibliometric data, citation networks, natural language processing, artificial intelligence, statistical physics, and models of collective memory to identify the conditions under which retracted research fades, persists, or spreads.

Temporal Characterization of Cultural Communities: A Computational Social Science Approach to Collective Memory and Attention

Mon, 30 Oct 2023 00:00:00 +0000

Collective forgetting is the process by which attention to cultural pieces decays over time. Research has shown that collective memory and attention decay in a universal two-step fashion, with a short high-intensive attention stage followed by a longer less-intensive attention stage. However, little is known about how this decay pattern changes over time. This project aims to study the temporal dynamics, changes, and effects of external shocks on collective memory and attention. The focus is on the differences between different domains of cultural production, such as knowledge and art, using a dynamic model and Big Data methods. The investigation will compare the attention dynamics of scientific articles and patents in the knowledge domain, and songs and films in the art domain. External shocks, such as health crises and awards, will also be studied. The study will contribute to understanding the dynamics of collective attention and formation of collective memories, and may have potential policy implications.

Send us an email to get more information about this project.

SocialRec: Personalizing Educational Experience through Social Network-based Recommender Systems

Mon, 01 Aug 2022 00:00:00 +0000

This project was undertaken as part of the University's Personalization Strategic Project, and it involved overcoming several challenges. The first of these was integrating data from various sources, each with their own unique formats, which required significant effort to ensure compatibility. Additionally, privacy and security were paramount concerns given the sensitive nature of the data. To address these issues, we developed both an anonymization protocol and a de-identification process that could preserve the necessary granularity of the data. Our team also developed a novel algorithm capable of creating a benchmark for inferring social networks from router connections. Finally, we integrated the analysis of mixed, individual, and relational data to build a recommendation system based on machine learning and interaction networks, which is a cutting-edge approach. Our work will be valuable for researchers and companies looking to develop state-of-the-art recommendation systems in their own research or business endeavors.

Send us an email to get more information about this project.

Lixandria: A Platform for discovering your degree program

Sun, 01 Dec 2019 00:00:00 +0000

Lixandria is a platform that offers guidance, visualization, and distribution of data on university careers. It utilizes the Higher Education Space (HES), which integrates information from historical applicants to the higher education system through their preferences and a network science-based algorithm. By doing so, the HES expands the options available to applicants when selecting their desired career. Additionally, Lixandria integrates and distributes data from multiple sources, such as the Ministry of Education, DEMRE, and mifuturo.cl, empowering applicants to the higher education system and decision-makers in higher education institutions and the central government.

The HES was originally developed as part of Cristian Candia’s Doctoral Thesis in 2018, which was carried out at the MIT media lab and the Center for Research in Social Complexity at UDD. The seminal article of Lixandria, which both constructs and validates the HES, was published in 2019 by Professor Cristian Candia (University of Desarrollo) in collaboration with Professor Flavio Pinheiro and researcher Sara Encarnaçao (both from Universidade NOVA de Lisboa).

Send us an email to get more information about this project.