UC San Diego

With the emergence of big and complex data collections, e.g., images, texts, and social networks, efficiently and effectively searching for the most relevant data samples with respect to a particular instance has become an important problem to investigate. For this purpose, this dissertation aims to design practical machine learning algorithms to efficiently and effectively learn to order in big and complex data so as to facilitate the underlying tasks such as link prediction/recommendation, visual search, recommender systems, or visualization. This study benefits both the theory and application of machine learning and has the potential to help domains such as data mining, computer vision, and bioinformatics.