Kernel methods enable the direct usage of structured representations of textual data during language learning and inference tasks. Expressive kernels, such as Tree Kernels, achieve excellent performance in NLP. On the other side, deep neural networks have been demonstrated effective in automatically learning feature representations during training. However, their input is tensor data, i.e., the...

Graph-structured data is important in modeling relationships between multiple entities, and can be used to represent states of the world as well as many data structures. Li et al. (2016) describe a model known as a Gated Graph Sequence Neural Network (GGS-NN) that produces sequences from graph-structured input. In this work I introduce the Gated Graph Transformer Neural Network (GGTNN), an exte...

Composition languages like BPEL and many enactment tools only support structured process models, while most composition approaches only consider unstructured models. In this paper, we outline a semi-automatic approach for composing a set of services with data flow dependencies into a structured process model. These data flow dependencies can be automatically derived from the input and output me...

Towards the goal of realizing a generic automatic human activity recognition system, a new formalism is proposed. Activities are described by a chained hierarchical representation using three type of entities: image features, mobile object properties and scenarios. Taking image features of tracked moving regions from an image sequence as input, mobile object properties are first computed by spe...

This paper 1 presents AutoLearn’s authoring tool: AutoTutor, a software solution that enables teachers (content creators) to develop language learning activities including automatic feedback generation without the need of being a programmer. The software has been designed and implemented on the basis of processing pipelines developed in previous work. A group of teachers has been trained to use...

Recognition of human motion provides hints to understand human activities and gives opportunities to the development of a new human-computer interaction (HCI) interface. Hidden Markov Models (HMMs) are used for visual recognition of complex, structured hand gestures such as the ones found in a sign language, since they have proved their success in recognizing speech and handwriting. In this pap...

The rehabilitation environment is structured differently from the hospital-based environment in a way that lends itself to interdisciplinary care. Physicians work with other specialists on an interprofessional team while observing patients' participation in activities of daily living. This approach allows a patient to show rather than tell a physician what he or she can do, which helps the phys...

The recent developments of statistical learning focused on vector machines, which learn from examples that are described by vectors of features. However, there are many fields where structured data must be handled; therefore, it would be desirable to learn from examples described by graphs. Graph machines learn real numbers from graphs. Basically, for each input graph, a separate learning machi...

One of the new design challenges in Information Retrieval (IR) and Human Computer Interaction (HCI) is to face a huge quantity of structured and accessible information which is available to a variety of users having different information needs. This article proposes a dimension space that should help in the design of current and future interactive information retrieval systems. To do so, we ado...

Understanding complex, dynamic scenes of real-world activities from low-level sensor data is of central importance for intelligent systems. The main difficulty lies in the fact that complex scenes are best described in high-level, logical formalisms, while sensor data usually consists of many low-level features. We first propose a method to obtain a logical representation of real-world, dynamic...