Full bibliographic details must be given when referring to, or quoting from full items including the author's name, the title of the work, publication details where relevant (place, publisher, date), pagination, and for theses or dissertations the awarding institution, the degree type awarded, and the date of the award.

The rapidly growing computational demands of deep neural networks require novel hardware designs. Recently, tunable nanoelectronic devices were developed based on hopping electrons through a network of dopant atoms in silicon. These "Dopant Network Processing Units" (DNPUs) are highly energy-efficient and have potentially very high throughput. By adapting the control voltages applied to its terminals, a single DNPU can solve a variety of linearly non-separable classification problems. However, using a single device has limitations due to the implicit single-node architecture. This paper presents a promising novel approach to neural information processing by introducing DNPUs as high-capacity neurons and moving from a single to a multi-neuron framework. By implementing and testing a small multi-DNPU classifier in hardware, we show that feed-forward DNPU networks improve the performance of a single DNPU from 77% to 94% test accuracy on a binary classification task with concentric classes on a plane. Furthermore, motivated by the integration of DNPUs with memristor arrays, we study the potential of using DNPUs in combination with linear layers. We show by simulation that a single-layer MNIST classifier with only 10 DNPUs achieves over 96% test accuracy. Our results pave the road towards hardware neural-network emulators that offer atomic-scale information processing with low latency and energy consumption.Preprint. Under review.

In terms of software engineering, context-aware systems (C-AS) have notably different development needs than those of traditional computing. Yet, there are no established methodologies that uniformly support the development life-cycle of these systems. A key goal of this research is to improve the current state-of-the-art with respect to engineering techniques for the life-cycle of a C-AS. Within the scope of this higher order goal, this paper addresses the lower level order goal of a holistic framework for gathering requirements which is specialised to the creation of C-AS. The framework follows an end-user, stakeholder-centred vision, which guides the analysis of stakeholders towards the discovery of specific stakeholder profiles and their particular needs, preferences, and limitations. It allows the operationalisation of the high level objectives of the system into requirements, which are more tangible and related to the implementation of the system. An evaluation procedure is supported, based on heuristics and rules from the NFR framework and REUBI. All the diagrams introduced for this framework have been developed as part of an open-source tool based on Modelio, which is intended to be developed in the future as part of a framework that covers all the stages of the development process. The proposal is illustrated through the analysis of an application for a European funded project.