Projected estimators for robust semi-supervised classification

For semi-supervised techniques to be applied safely in practice we at least want methods to outperform their supervised counterparts. We study this question for classification using the well-known quadratic surrogate loss function. Unlike other approaches to semi-supervised learning, the procedure proposed in this work does not rely on...

Design of a novel presentation product: Whitetable: a portable interactive storytelling tool

This thesis describes the design process of a novel portable interactive storytelling tool for the corporate presentation market. The tool allows the user to tell complex stories in a clear and engaging manner by relieving the restrictions found in linear presentation tools such as slide decks. The product consists of a tabletop downwards firing...

Utilizing a Discrete Global Grid System For Handling Point Clouds With Varying Locations, Times, and Levels of Detail

Discrete Global Grid Systems (DGGS) have emerged in recent years as a new specification for working with global heterogeneous datasets of different types, such as vector, raster, and quantitative datasets. There exists no research on using point clouds with a DGGS. This research aims to analyse the extent to which a DGGS can be used to handle...

Pressure and Velocity Fields Retrieval Based on a New Normalization Option and 1D Model-Free Primaries Retrieval from Marine Data: New Marchenko-type Equations with Two-Way Wavefields at the Receiver Level

The Marchenko method is being developed to compute redatuming operators from the reflection response based on a new normalization option at the data level. These operators can be used to retrieve the two-way focusing functions in terms of pressure and velocity, respectively, without effects from the ghost, and free-surface and internal multiples...