this paper presents an accurate differential global positioning system (dgps) using multi-layered neural networks (nns) based on the back propagation (bp) and imperialistic competition algorithm (ica) in order to predict the dgps corrections for accurate positioning. simulation results allowed us to optimize the nn performance in term of residual mean square error. we compare results obtained b...

Communications between systems is often initially represented at a single, high level of abstraction, a link between components. During design evolution it is usually necessary to elaborate the interface model, defining it from several different, related viewpoints and levels of abstraction. This paper presents a pattern to model such multi-layered interface architectures simply and efficiently...

In this project, we investigate the feasibility of using outdated machines with slow processors for tolerating disk latencies for computation and data intensive parallel adaptive and irregular applications.

INTRODUCTION Today, users have access to a multitude of interactive systems, such as desktop computers, cell phones, and interactive whiteboards. However, all these systems are designed to be used mostly independently and are severely limited in their cross-system capabilities: applications are encapsulated by the system they run on and cannot easily transfer or extend their user interface (UI)...

This paper discusses the stackable layers approach to le system design. With this approach, a le system is constructed from several layers, each implementing one portion of the le system well. Each layer is bounded above and below by an identical interface framework. The symmetry of the interface, coupled with run-time stack deenition, make layer conngura-tion exible and facilitate experimentat...

The author’s original state-space learning system (based on a probabilistic performance measure clustered in feature space) was effective in optimizing parameterized linear evaluation functions. However, more accurate probability estimates would allow stabilization in cases of strong feature interactions. To attain this accuracy and stability, a second level of learning is added, a genetic (par...

A typical NLP system for medical fact coding uses multiple layers of supervision involving factattributes, relations and coding. Training such a system involves expensive and laborious annotation process involving all layers of the pipeline. In this work, we investigate the feasibility of a shallow medical coding model that trains only on fact annotations, while disregarding fact-attributes and...

Abstract. In this paper, a hierarchical system architecture for singleagent multi-modal systems is proposed. The layered system is designed to promote proof obligations so that system specification at one level of granularity conforms with that at another level and vice versa. The design principle for the construction of the hierarchy is based on bisimulation with respect to reachability specif...

Public Good Software’s products match journalistic articles and other narrative content to relevant charitable causes and nonprofit organizations so that readers can take action on the issues raised by the articles’ publishers. Previously an expensive and laborintensive process, application of machine learning and other automated textual analyses now allow us to scale this matching process to t...