Descripción del proyecto
Nuevos métodos estadísticos para reducir la incertidumbre en el análisis de datos masivos
Los datos masivos suelen ser confusos: la fracción de errores en los datos crece conforme aumenta el volumen de datos. Obtener conclusiones fiables puede resultar complicado cuando los datos se recopilan en diferentes condiciones, o cuando algunos datos faltan o están dañados. El objetivo del proyecto RobustStats, financiado con fondos europeos, es desarrollar una metodología y una teoría estadísticas sólidas para abordar los retos de los datos masivos. En el contexto del aprendizaje por transferencia, sus investigadores emplearán los métodos adecuados para explotar las distribuciones entre los dominios de origen y de destino. Además, probarán los mecanismos de datos perdidos y desarrollarán herramientas prácticas para manejar tanto datos perdidos como datos heterogéneos en las etiquetas de clasificación. En último término, se introducirá la perturbación de datos para lograr una inferencia sólida con datos masivos.
Objetivo
Modern technology allows large-scale data to be collected in many new forms, and their underlying generating mechanisms can be extremely complex. In fact, an interesting (and perhaps initially surprising) feature of large-scale data is that it is often much harder to feel confident that one has identified a plausible statistical model. This is largely because there are so many forms of model violation and both visual and more formal statistical checks can become infeasible. It is therefore vital for trust in conclusions drawn from large studies that statisticians ensure that their methods are robust. The RobustStats proposal will introduce new statistical methodology and theory for a range of important contemporary Big Data challenges. In transfer learning, we wish to make inference about a target data population, but some (typically, most) of our training data come from a related but distinct source distribution. The central goal is to find appropriate ways to exploit the relationship between the source and target distributions. Missing and corrupted data play an ever more prominent role in large-scale data sets because the proportion of cases with no missing attributes is typically small. We will address key challenges of testing the form of the missingness mechanism, and handling heterogeneous missingness and corruptions in classification labels. The robustness of a statistical procedure is intimately linked to model misspecification. We will advocate for two approaches to studying model misspecification, one via the idea of regarding an estimator as a projection onto a model, and the other via oracle inequalities. Finally, we will introduce new methods for robust inference with large-scale data based on the idea of data perturbation. Such approaches are attractive ways of exploring a space of distributions in a model-free way, and we will show that aggregation of the results of carefully-selected perturbations can be highly effective.
Ámbito científico
Programa(s)
Régimen de financiación
ERC-ADG - Advanced GrantInstitución de acogida
CB2 1TN Cambridge
Reino Unido