Large-scale kernel ridge regression (KRR) is limited by the need to store a large kernel matrix Kt. To avoid storing the entire matrix Kt, Nyström methods subsample a subset of columns of the kernel matrix, and efficiently find an approximate KRR solution on the reconstructed K̃t. The chosen subsampling distribution in turn affects the statistical and computational tradeoffs. For KRR problems, [...

Recent studies show that large-scale sketch-based image retrieval (SBIR) can be efficiently tackled by cross-modal binary representation learning methods, where Hamming distance matching significantly speeds up the process of similarity search. Providing training and test data subjected to a fixed set of pre-defined categories, the cutting-edge SBIR and cross-modal hashing works obtain acceptab...

Abstract. In the Corpus Architect project we are developing a system for creating text corpora, which will be easy to use even for non-technical users. The system creates corpora from two data sources – users’ own text documents and web pages. Users can upload their texts in various formats or ask the system for adding domain specific web texts into their corpora in an automated way. In the lat...

We present an image retrieval system for the interactive search of photo collections using free-hand sketches depicting shape. We describe Gradient Field HOG (GF-HOG); an adapted form of the HOG descriptor suitable for sketch based image retrieval (SBIR). We incorporate GF-HOG into a Bag of Visual Words (BoVW) retrieval framework, and demonstrate how this combination may be harnessed both for r...

Sharing feature-based computer-aided design (CAD) models is a challenging problem that is frequently encountered among heterogeneous CAD systems. In this work, a new asymmetric strategy is presented to enrich the theory of feature-based interoperability, particularly when addressing a singular feature or singular sketch. This paper analyzes the semantic asymmetry singular feature interoperabili...