
handle: 11441/77573
En este trabajo presentamos los conceptos principales sobre privacidad en Ciencia de Datos. Introducimos la definición de Privacidad Diferencial, que se ha desarrollado en los últimos años como una noción fundamental para la investigación en esta materia. Desarrollamos sus propiedades y hacemos un estudio introductorio y autocontenido de dos herramientas clave: el mecanismo de Laplace y el mecanismo exponencial. Ofrecemos también un caso práctico donde vemos la teoría en ejecución; exponemos la Técnica del Vector Disperso como un método con abundantes usos en el análisis de datos y terminamos mencionando dos librerías con el código informático oportuno para aproximarnos a estas cuestiones.
This final project shows the main concepts related to privacy in Data Science. We present the definition of Differential Privacy (DP), which has emerged as the standard privacy notion for research in this topic, and focus on some of its key properties. Then, we explore the Laplacian Mechanism and the Exponential Mechanism as two fundamental tools for achieving DP, i.e. strong guarantee against an open world environment. We also give a case study about these questions and provide a theoretical analysis of the Sparse Vector Technique, presenting some experimental libraries to see how its works in detail.
Universidad de Sevilla. Grado en Matemáticas
Privacidad diferencial, Ciencia de los datos
Privacidad diferencial, Ciencia de los datos
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 0 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Average | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
