IMFD researchers awarded Fondecyt Regular 2026 projects
IMFD researchers and academics from the Department of Computer Science at the University of Chile, Benjamin Bustos and Gonzalo Navarro, were each awarded a Fondecyt Regular project in the 2026 call for proposals. These projects address time- and space-efficient techniques for extending worst-case optimal algorithms to complex graph and relational databases, and a multi-modal similarity model that integrates text, images, and 3D models to improve the search for ancient ceramics in digital collections.
The Fondecyt project, administered by the National Research and Development Agency (ANID), aims to promote scientific and technological research in various areas of knowledge by funding individual research projects of excellence focused on knowledge production.
The awarded projects
Researcher Gonzalo Navarro was awarded the Space/Time Efficient Solutions to Extend Graph Databases project, which seeks to develop new techniques for relational and graph databases that improve solutions based on worst-case optimal (WCO) algorithms, extending them to practical scenarios to increase their usefulness and accessibility.
The research will focus, on the one hand, on optimizing fundamental queries through better data structures, such as hashing and factorized representations, and on the other hand, on expanding these algorithms to support more complex models, including Property Graphs and temporal or spatial data, thus bridging the gap between research prototypes and professional use. We will also seek to generalize these results to the relational model, integrating WCO algorithms with classic join techniques to improve space efficiency. By extending the solutions to tables with more attributes and combining compact solutions with binary joins and factorization, we hope that the use of WCO algorithms will become more frequent and convenient in practice.
On the other hand, Benjamín Bustos will work on the Multi-Modal Retrieval for Ancient Pottery research project. The project focuses on searching collections of ancient pottery. The challenges in solving this problem stem from the context of the data: sources of this type of data are scarce; objects may be incomplete, with eroded parts, or may contain noise or errors generated during the digitization process; these objects may be described using a combination of textual information, images, and 3D geometry.
The problem is how to characterize ancient ceramics in a way that represents all modalities simultaneously (text, image, 3D model) and allows searches for similar objects. "In this project, we propose to design a multi-modal similarity model in the context of ancient ceramics, which will allow us to implement search algorithms for multi-modal queries. The idea is that, using our multi-modal similarity model, queries can be made that combine the different modalities, for example, searching for objects in the collection that have the best match given a text and a reference image," says the researcher.
Every year, the National Fund for Scientific and Technological Development (Fondecyt) encourages and promotes the development of basic scientific and technological research. It has funded more than 16,000 research projects whose impacts have benefited both the scientific community and society in general.
Sources: DCC University of Chile