The critical challenge of single image inpainting stems from accurate semantic inference via limited information while maintaining quality. Typical methods for train an encoder-decoder network by learning a one-to-one mapping the corrupted to inpainted version. While such perform well on images with small regions, it is challenging these deal large area due two potential limitations. 1) Such pa...

Relation extraction tasks aim to predict potential relations between entities in a target sentence. As entity mentions have ambiguity sentences, some important contextual information can guide the semantic representation of improve accuracy relation extraction. However, most existing models ignore guidance and treat textual context sentence equally. This results low-accuracy extractions. To add...

Producing densely annotated data is a difficult and tedious task for medical imaging applications. To address this problem, we propose novel approach to generate supervision semi-supervised semantic segmentation. We argue that visually similar regions between labeled unlabeled images likely contain the same semantics therefore should share their label. Following thought, use small number of as ...

When facilitating interoperability at the data level one faces the problem that different data models are used as the basis for business formats. For example relational databases are based on the relational model, while XML Schema is basically a hierarchical model (with some extensions, like references). Our goal is to provide a syntax and a data model neutral format for the representation of b...

Zero-Shot Learning (ZSL) promises to scale visual recognition by bypassing the conventional model training requirement of annotated examples for every category. This is achieved by establishing a mapping connecting low-level features and a semantic description of the label space, referred as visual-semantic mapping, on auxiliary data. Reusing the learned mapping to project target videos into an...

This paper addresses the problem of e cient management of events, in particular in those environments where events carry information useful to multiple applications, possibly operating in di erent domains and at di erent levels of abstraction. We investigate the problems and opportunities o ered by such environments, and de ne a framework that enables a semantic mapping of events, i.e., enables...

Masters, Mori, & Mori (1993) discuss the use of semantic mapping strategy. They define semantic mapping technique as being "used to motivate and involve students in the thinking, reading, and writing aspects. It enhances vocabulary development by helping students link new information with previous experience." The instructional sequence of semantic mapping is as follows: 1. Select a word centra...

To survive in the twenty-first century, enterprises need to collaborate. Collaboration at the enterprise-level presupposes the interoperability of the underlying information systems. Access to heterogeneous information sources must be provided transparently while maintaining their autonomy. Further, the availability of nearly unlimited information calls for efficient and precise information ret...