The recommended research to be awarded in the present process of this contest are:
"Semantic Optimization of Conjunctive Queries in Treatable Classes" by Pablo Barceló.
The conjunctive consultations in databases are among the most popular in practice, and also the most studied. However, evaluating such queries is difficult, particularly in the current scenario where the volumes of data being queried become increasingly large. Our project focuses on understanding how to extract the maximum information about the structure of the data in order to alleviate the process of evaluation of conjunctive consultations. In particular, we study how this information can be used to rewrite the query as another for which we have theoretical assurances about the efficiency of its evaluation problem.
"Static Analysis and Learning Problems for JSON Schema" by John L. Reutter
The world of the Web is dominated by APIs: services to which all audiences can connect to extract information. The vast majority of services we know have APIs (like Google, Facebook and Twitter). Using APIs we can access all these services at the same time, integrating and complementing the information that each of them provides. The problem is that today it is very difficult to work with several APIs, because each of them has its own way of working and exposing data. There is currently an initiative to achieve a standard - a common language in which all API providers agree to describe what they provide and how their APIs work. In this context, my research has to do with an automatic way of learning the main characteristics of Web APIs, and in particular the structure of the files they expect as inputs and outputs. The development of these algorithms would allow us to bring the documentation available in each API to a common format, which would facilitate the adoption of the standard and bring us closer to the dream of a fully integrated Web.