In this paper, we describe a retrieval system that uses hidden annotation to improve the performance. The contribution of this paper is a novel active learning framework that can ...
Learning effectiveness depends on a large range of parameters. Learners’ activity has an important impact on long-term learning and comprehension of difficult concepts [1]. Coll...
Labeling text data is quite time-consuming but essential for automatic text classification. Especially, manually creating multiple labels for each document may become impractical ...
Model selection is important in many areas of supervised learning. Given a dataset and a set of models for predicting with that dataset, we must choose the model which is expected...
Supervised learning of a parts-based model can be for-
mulated as an optimization problem with a large (exponen-
tial in the number of parts) set of constraints. We show how
thi...
M. Pawan Kumar, Andrew Zisserman, Philip H.S. Torr