Abstract. Conceptual modelling tools allow users to construct formal representations of their conceptualisations. These models are typically developed in isolation, unrelated to ot...
This paper revisits the pivot language approach for machine translation. First, we investigate three different methods for pivot translation. Then we employ a hybrid method combin...
Stochastic gradient descent (SGD) uses approximate gradients estimated from subsets of the training data and updates the parameters in an online fashion. This learning framework i...
The varying object appearance and unlabeled data from new frames are always the challenging problem in object tracking. Recently machine learning methods are widely applied to tra...
We propose a new method for handwritten word-spotting which does not require prior training or gathering examples for querying. More precisely, a model is trained "on the fly...