A Bayesian perspective on stochastic neurocontrol

Control design for stochastic uncertain nonlinear systems is traditionally based on minimizing the expected value of a suitably chosen loss function. Moreover, most control methods usually assume the certainty equivalence principle to simplify the problem and make it computationally tractable. We offer an improved probabilistic framework which is not constrained by these previous assumptions, and provides a more natural framework for incorporating and dealing with uncertainty. The focus of this paper is on developing this framework to obtain an optimal control law strategy using a fully probabilistic approach for information extraction from process data, which does not require detailed knowledge of system dynamics. Moreover, the proposed control method framework allows handling the problem of input-dependent noise. A basic paradigm is proposed and the resulting algorithm is discussed. The proposed probabilistic control method is for the general nonlinear class of discrete-time systems. It is demonstrated theoretically on the affine class. A nonlinear simulation example is also provided to validate theoretical development.

Stochastic beamforming for cochlear implant coding

Cochlear implants are prosthetic devices used to provide hearing to people who would otherwise be profoundly deaf. The deliberate addition of noise to the electrode signals could increase the amount of information transmitted, but standard cochlear implants do not replicate the noise characteristic of normal hearing because if noise is added in an uncontrolled manner with a limited number of electrodes then it will almost certainly lead to worse performance. Only if partially independent stochastic activity can be achieved in each nerve fibre can mechanisms like suprathreshold stochastic resonance be effective. We are investigating the use of stochastic beamforming to achieve greater independence. The strategy involves presenting each electrode with a linear combination of independent Gaussian noise sources. Because the cochlea is filled with conductive salt solutions, the noise currents from the electrodes interact and the effective stimulus for each nerve fibre will therefore be a different weighted sum of the noise sources. To some extent therefore, the effective stimulus for a nerve fibre will be independent of the effective stimulus of neighbouring fibres. For a particular patient, the electrode position and the amount of current spread are fixed. The objective is therefore to find the linear combination of noise sources that leads to the greatest independence between nerve discharges. In this theoretical study we show that it is possible to get one independent point of excitation (one null) for each electrode and that stochastic beamforming can greatly decrease the correlation between the noise exciting different regions of the cochlea. © 2007 Copyright SPIE - The International Society for Optical Engineering.

Requirements-aware systems:a research agenda for RE for self-adaptive systems

Requirements are sensitive to the context in which the system-to-be must operate. Where such context is well-understood and is static or evolves slowly, existing RE techniques can be made to work well. Increasingly, however, development projects are being challenged to build systems to operate in contexts that are volatile over short periods in ways that are imperfectly understood. Such systems need to be able to adapt to new environmental contexts dynamically, but the contextual uncertainty that demands this self-adaptive ability makes it hard to formulate, validate and manage their requirements. Different contexts may demand different requirements trade-offs. Unanticipated contexts may even lead to entirely new requirements. To help counter this uncertainty, we argue that requirements for self-adaptive systems should be run-time entities that can be reasoned over in order to understand the extent to which they are being satisfied and to support adaptation decisions that can take advantage of the systems' self-adaptive machinery. We take our inspiration from the fact that explicit, abstract representations of software architectures used to be considered design-time-only entities but computational reflection showed that architectural concerns could be represented at run-time too, helping systems to dynamically reconfigure themselves according to changing context. We propose to use analogous mechanisms to achieve requirements reflection. In this paper we discuss the ideas that support requirements reflection as a means to articulate some of the outstanding research challenges.

Goal-based modeling of dynamically adaptive system requirements

Self-adaptation is emerging as an increasingly important capability for many applications, particularly those deployed in dynamically changing environments, such as ecosystem monitoring and disaster management. One key challenge posed by Dynamically Adaptive Systems (DASs) is the need to handle changes to the requirements and corresponding behavior of a DAS in response to varying environmental conditions. Berry et al. previously identified four levels of RE that should be performed for a DAS. In this paper, we propose the Levels of RE for Modeling that reify the original levels to describe RE modeling work done by DAS developers. Specifically, we identify four types of developers: the system developer, the adaptation scenario developer, the adaptation infrastructure developer, and the DAS research community. Each level corresponds to the work of a different type of developer to construct goal model(s) specifying their requirements. We then leverage the Levels of RE for Modeling to propose two complementary processes for performing RE for a DAS. We describe our experiences with applying this approach to GridStix, an adaptive flood warning system, deployed to monitor the River Ribble in Yorkshire, England.

Mitigating the obsolescence of quality specifications models in service-based systems

Requirements-aware systems address the need to reason about uncertainty at runtime to support adaptation decisions, by representing quality of services (QoS) requirements for service-based systems (SBS) with precise values in run-time queryable model specification. However, current approaches do not support updating of the specification to reflect changes in the service market, like newly available services or improved QoS of existing ones. Thus, even if the specification models reflect design-time acceptable requirements they may become obsolete and miss opportunities for system improvement by self-adaptation. This articles proposes to distinguish "abstract" and "concrete" specification models: the former consists of linguistic variables (e.g. "fast") agreed upon at design time, and the latter consists of precise numeric values (e.g. "2ms") that are dynamically calculated at run-time, thus incorporating up-to-date QoS information. If and when freshly calculated concrete specifications are not satisfied anymore by the current service configuration, an adaptation is triggered. The approach was validated using four simulated SBS that use services from a previously published, real-world dataset; in all cases, the system was able to detect unsatisfied requirements at run-time and trigger suitable adaptations. Ongoing work focuses on policies to determine recalculation of specifications. This approach will allow engineers to build SBS that can be protected against market-caused obsolescence of their requirements specifications. © 2012 IEEE.

Tracing requirements for adaptive systems using claims

The complexity of environments faced by dynamically adaptive systems (DAS) means that the RE process will often be iterative with analysts revisiting the system specifications based on new environmental understanding product of experiences with experimental deployments, or even after final deployments. An ability to trace backwards to an identified environmental assumption, and to trace forwards to find the areas of a DAS's specification that are affected by changes in environmental understanding aids in supporting this necessarily iterative RE process. This paper demonstrates how claims can be used as markers for areas of uncertainty in a DAS specification. The paper demonstrates backward tracing using claims to identify faulty environmental understanding, and forward tracing to allow generation of new behaviour in the form of policy adaptations and models for transitioning the running system. © 2011 ACM.

Dynamic decision networks for decision-making in self-adaptive systems:a case study

Bayesian decision theory is increasingly applied to support decision-making processes under environmental variability and uncertainty. Researchers from application areas like psychology and biomedicine have applied these techniques successfully. However, in the area of software engineering and speci?cally in the area of self-adaptive systems (SASs), little progress has been made in the application of Bayesian decision theory. We believe that techniques based on Bayesian Networks (BNs) are useful for systems that dynamically adapt themselves at runtime to a changing environment, which is usually uncertain. In this paper, we discuss the case for the use of BNs, speci?cally Dynamic Decision Networks (DDNs), to support the decision-making of self-adaptive systems. We present how such a probabilistic model can be used to support the decision making in SASs and justify its applicability. We have applied our DDN-based approach to the case of an adaptive remote data mirroring system. We discuss results, implications and potential bene?ts of the DDN to enhance the development and operation of self-adaptive systems, by providing mechanisms to cope with uncertainty and automatically make the best decision.

Architecting self-aware software systems

Contemporary software systems are becoming increasingly large, heterogeneous, and decentralised. They operate in dynamic environments and their architectures exhibit complex trade-offs across dimensions of goals, time, and interaction, which emerges internally from the systems and externally from their environment. This gives rise to the vision of self-aware architecture, where design decisions and execution strategies for these concerns are dynamically analysed and seamlessly managed at run-time. Drawing on the concept of self-awareness from psychology, this paper extends the foundation of software architecture styles for self-adaptive systems to arrive at a new principled approach for architecting self-aware systems. We demonstrate the added value and applicability of the approach in the context of service provisioning to cloud-reliant service-based applications.

Bayesian precalibration of a large stochastic microsimulation model

Calibration of stochastic traffic microsimulation models is a challenging task. This paper proposes a fast iterative probabilistic precalibration framework and demonstrates how it can be successfully applied to a real-world traffic simulation model of a section of the M40 motorway and its surrounding area in the U.K. The efficiency of the method stems from the use of emulators of the stochastic microsimulator, which provides fast surrogates of the traffic model. The use of emulators minimizes the number of microsimulator runs required, and the emulators' probabilistic construction allows for the consideration of the extra uncertainty introduced by the approximation. It is shown that automatic precalibration of this real-world microsimulator, using turn-count observational data, is possible, considering all parameters at once, and that this precalibrated microsimulator improves on the fit to observations compared with the traditional expertly tuned microsimulation. © 2000-2011 IEEE.

Probabilistic control for uncertain systems

In this paper a new framework has been applied to the design of controllers which encompasses nonlinearity, hysteresis and arbitrary density functions of forward models and inverse controllers. Using mixture density networks, the probabilistic models of both the forward and inverse dynamics are estimated such that they are dependent on the state and the control input. The optimal control strategy is then derived which minimizes uncertainty of the closed loop system. In the absence of reliable plant models, the proposed control algorithm incorporates uncertainties in model parameters, observations, and latent processes. The local stability of the closed loop system has been established. The efficacy of the control algorithm is demonstrated on two nonlinear stochastic control examples with additive and multiplicative noise.

Improving scheduling techniques in heterogeneous systems with dynamic, on-line optimisations

Computational performance increasingly depends on parallelism, and many systems rely on heterogeneous resources such as GPUs and FPGAs to accelerate computationally intensive applications. However, implementations for such heterogeneous systems are often hand-crafted and optimised to one computation scenario, and it can be challenging to maintain high performance when application parameters change. In this paper, we demonstrate that machine learning can help to dynamically choose parameters for task scheduling and load-balancing based on changing characteristics of the incoming workload. We use a financial option pricing application as a case study. We propose a simulation of processing financial tasks on a heterogeneous system with GPUs and FPGAs, and show how dynamic, on-line optimisations could improve such a system. We compare on-line and batch processing algorithms, and we also consider cases with no dynamic optimisations.

Quality Estimation of e-Learning Systems

Existing approaches to quality estimation of e-learning systems are analyzed. The “layered” approach for quality estimation of e-learning systems enhanced with learning process modeling and simulation is presented. The method of quality estimation using learning process modeling and quality criteria are suggested. The learning process model based on extended colored stochastic Petri net is described. The method has been implemented in the automated system of quality estimation of e-learning systems named “QuAdS”. Results of approbation of the developed method and quality criteria are shown. We argue that using learning process modeling for quality estimation simplifies identifying lacks of an e-learning system for an expert.

Stochastic Arithmetic Theory and Experiments

Stochastic arithmetic has been developed as a model for exact computing with imprecise data. Stochastic arithmetic provides confidence intervals for the numerical results and can be implemented in any existing numerical software by redefining types of the variables and overloading the operators on them. Here some properties of stochastic arithmetic are further investigated and applied to the computation of inner products and the solution to linear systems. Several numerical experiments are performed showing the efficiency of the proposed approach.

Architecture of Extensible Computations Driven Systems

Никола Вълчанов, Тодорка Терзиева, Владимир Шкуртов, Антон Илиев - Една от основните области на приложения на компютърната информатика е автоматизирането на математическите изчисления. Информационните системи покриват различни области като счетоводство, електронно обучение/тестване, симулационни среди и т. н. Те работят с изчислителни библиотеки, които са специфични за обхвата на системата. Въпреки, че такива системи са перфектни и работят безпогрешно, ако не се поддържат остаряват. В тази работа описваме механизъм, който използва динамично библиотеките за изчисления и взема решение по време на изпълнение (интелигентно или интерактивно) за това как и кога те да се използват. Целта на тази статия е представяне на архитектура за системи, управлявани от изчисления. Тя се фокусира върху ползите от използването на правилните шаблони за дизайн с цел да се осигури разширяемост и намаляване на сложността.

Generalized Discernibility Function Based Attribute Reduction in Incomplete Decision Systems

A rough set approach for attribute reduction is an important research subject in data mining and machine learning. However, most attribute reduction methods are performed on a complete decision system table. In this paper, we propose methods for attribute reduction in static incomplete decision systems and dynamic incomplete decision systems with dynamically-increasing and decreasing conditional attributes. Our methods use generalized discernibility matrix and function in tolerance-based rough sets.