Streaming covariance selection with applications to adaptive querying in sensor networks

Sensor networks can be naturally represented as graphical models, where the edge set encodes the presence of sparsity in the correlation structure between sensors. Such graphical representations can be valuable for information mining purposes as well as for optimizing bandwidth and battery usage with minimal loss of estimation accuracy. We use a computationally efficient technique for estimating sparse graphical models which fits a sparse linear regression locally at each node of the graph via the Lasso estimator. Using a recently suggested online, temporally adaptive implementation of the Lasso, we propose an algorithm for streaming graphical model selection over sensor networks. With battery consumption minimization applications in mind, we use this algorithm as the basis of an adaptive querying scheme. We discuss implementation issues in the context of environmental monitoring using sensor networks, where the objective is short-term forecasting of local wind direction. The algorithm is tested against real UK weather data and conclusions are drawn about certain tradeoffs inherent in decentralized sensor networks data analysis. © 2010 The Author. Published by Oxford University Press on behalf of The British Computer Society. All rights reserved.

On applications of high-frequency asymptotics in aeroacoustics.

The aim of this paper is to survey a range of applications of high-frequency asymptotic methods in aeroacoustics. Specifically, we are concerned with problems associated with noise generation, propagation and scattering as found in large modern aeroengines. With regard to noise generation, we consider the interaction between high-frequency vortical waves and thin aerofoils, with particular emphasis being placed on the way in which the vortical waves act on the non-uniform mean flow around the aerofoil. A ray-theoretic description of the resulting sound as it propagates along the engine intake is then presented, followed by consideration of the diffraction of these rays by the (possibly asymmetric) intake lip to produce sound in the far field. A range of more detailed possible extensions is also presented.

Simulating dynamic covariance structures for testing the adaptive behavior of variable selection algorithms (Invited Paper)

Variable selection for regression is a classical statistical problem, motivated by concerns that too large a number of covariates may bring about overfitting and unnecessarily high measurement costs. Novel difficulties arise in streaming contexts, where the correlation structure of the process may be drifting, in which case it must be constantly tracked so that selections may be revised accordingly. A particularly interesting phenomenon is that non-selected covariates become missing variables, inducing bias on subsequent decisions. This raises an intricate exploration-exploitation tradeoff, whose dependence on the covariance tracking algorithm and the choice of variable selection scheme is too complex to be dealt with analytically. We hence capitalise on the strength of simulations to explore this problem, taking the opportunity to tackle the difficult task of simulating dynamic correlation structures. © 2008 IEEE.

Temporally adaptive estimation of logistic classifiers on data streams

Modern technology has allowed real-time data collection in a variety of domains, ranging from environmental monitoring to healthcare. Consequently, there is a growing need for algorithms capable of performing inferential tasks in an online manner, continuously revising their estimates to reflect the current status of the underlying process. In particular, we are interested in constructing online and temporally adaptive classifiers capable of handling the possibly drifting decision boundaries arising in streaming environments. We first make a quadratic approximation to the log-likelihood that yields a recursive algorithm for fitting logistic regression online. We then suggest a novel way of equipping this framework with self-tuning forgetting factors. The resulting scheme is capable of tracking changes in the underlying probability distribution, adapting the decision boundary appropriately and hence maintaining high classification accuracy in dynamic or unstable environments. We demonstrate the scheme's effectiveness in both real and simulated streaming environments. © Springer-Verlag 2009.

Collective Cell Migration: Leadership, Invasion and Segregation

A number of biological processes, such as embryo development, cancer metastasis or wound healing, rely on cells moving in concert. The mechanisms leading to the emergence of coordinated motion remain however largely unexplored. Although biomolecular signalling is known to be involved in most occurrences of collective migration, the role of physical and mechanical interactions has only been recently investigated. In this paper, a versatile framework for cell motility is implemented in-silico in order to study the minimal requirements for the coordination of a group of epithelial cells. We find that cell motility and cell-cell mechanical interactions are sufficient to generate a broad array of behaviours commonly observed in vitro and in vivo. Cell streaming, sheet migration and susceptibility to leader cells are examples of behaviours spontaneously emerging from these simple assumptions, which might explain why collective effects are so ubiquitous in nature. This analysis provides also new insights into cancer metastasis and cell sorting, suggesting in particular that collective invasion might result from an emerging coordination in a system where single cells are mechanically unable to invade.

An experimental model of relay deployment planning tool for a wireless sensor network system to monitor civil engineering structures

It is essential to monitor deteriorated civil engineering structures cautiously to detect symptoms of their serious disruptions. A wireless sensor network can be an effective system for monitoring civil engineering structures. It is fast to deploy sensors especially in difficult-to-access areas, and it is extendable without any cable extensions. Since our target is to monitor deteriorations of civil engineering structures such as cracks at tunnel linings, most of the locations of sensors are known, and sensors are not required to move dynamically. Therefore, we focus on developing a deployment plan of a static network in order to reduce the value of a cost function such as initial installation cost and summation of communication distances of the network. The key issue of the deployment is the location of relays that forward sensing data from sensors to a data collection device called a gateway. In this paper, we propose a relay deployment-planning tool that can be used to design a wireless sensor network for monitoring civil engineering structures. For the planning tool, we formalize the model and implement a local search based algorithm to find a quasi-optimal solution. Our solution guarantees two routings from a sensor to a gateway, which can provide higher reliability of the network. We also show the application of our experimental tool to the actual environment in the London Underground.