Towards optimizing the selection of neurons in affordable neural networks

This paper considers variations of a neuron pool selection method known as Affordable Neural Network (AfNN). A saliency measure, based on the second derivative of the objective function is proposed to assess the ability of a trained AfNN to provide neuronal redundancy. The discrepancies between the various affordability variants are explained by correlating unique sub group selections with relevant saliency variations. Overall this study shows that the method in which neurons are selected from a pool is more relevant to how salient individual neurons are, than how often a particular neuron is used during training. The findings herein are relevant to not only providing an analogy to brain function but, also, in optimizing the way a neural network using the affordability method is trained.

A peak reduction scheduling algorithm for storage devices on the low voltage network

Reinforcing the Low Voltage (LV) distribution network will become essential to ensure it remains within its operating constraints as demand on the network increases. The deployment of energy storage in the distribution network provides an alternative to conventional reinforcement. This paper presents a control methodology for energy storage to reduce peak demand in a distribution network based on day-ahead demand forecasts and historical demand data. The control methodology pre-processes the forecast data prior to a planning phase to build in resilience to the inevitable errors between the forecasted and actual demand. The algorithm uses no real time adjustment so has an economical advantage over traditional storage control algorithms. Results show that peak demand on a single phase of a feeder can be reduced even when there are differences between the forecasted and the actual demand. In particular, results are presented that demonstrate when the algorithm is applied to a large number of single phase demand aggregations that it is possible to identify which of these aggregations are the most suitable candidates for the control methodology.

Gaussian processes autoencoder for dimensionality reduction

Learning low dimensional manifold from highly nonlinear data of high dimensionality has become increasingly important for discovering intrinsic representation that can be utilized for data visualization and preprocessing. The autoencoder is a powerful dimensionality reduction technique based on minimizing reconstruction error, and it has regained popularity because it has been efficiently used for greedy pretraining of deep neural networks. Compared to Neural Network (NN), the superiority of Gaussian Process (GP) has been shown in model inference, optimization and performance. GP has been successfully applied in nonlinear Dimensionality Reduction (DR) algorithms, such as Gaussian Process Latent Variable Model (GPLVM). In this paper we propose the Gaussian Processes Autoencoder Model (GPAM) for dimensionality reduction by extending the classic NN based autoencoder to GP based autoencoder. More interestingly, the novel model can also be viewed as back constrained GPLVM (BC-GPLVM) where the back constraint smooth function is represented by a GP. Experiments verify the performance of the newly proposed model.

Acceleration and visualization of Dynamic Network Optimization

With the emerging prevalence of smart phones and 4G LTE networks, the demand for faster-better-cheaper mobile services anytime and anywhere is ever growing. The Dynamic Network Optimization (DNO) concept emerged as a solution that optimally and continuously tunes the network settings, in response to varying network conditions and subscriber needs. Yet, the DNO realization is still at infancy, largely hindered by the bottleneck of the lengthy optimization runtime. This paper presents the design and prototype of a novel cloud based parallel solution that further enhances the scalability of our prior work on various parallel solutions that accelerate network optimization algorithms. The solution aims to satisfy the high performance required by DNO, preliminarily on a sub-hourly basis. The paper subsequently visualizes a design and a full cycle of a DNO system. A set of potential solutions to large network and real-time DNO are also proposed. Overall, this work creates a breakthrough towards the realization of DNO.

Collocating satellite-based radar and radiometer measurements — methodology and usage examples

Collocations between two satellite sensors are occasions where both sensors observe the same place at roughly the same time. We study collocations between the Microwave Humidity Sounder (MHS) on-board NOAA-18 and the Cloud Profiling Radar (CPR) on-board CloudSat. First, a simple method is presented to obtain those collocations and this method is compared with a more complicated approach found in literature. We present the statistical properties of the collocations, with particular attention to the effects of the differences in footprint size. For 2007, we find approximately two and a half million MHS measurements with CPR pixels close to their centrepoints. Most of those collocations contain at least ten CloudSat pixels and image relatively homogeneous scenes. In the second part, we present three possible applications for the collocations. Firstly, we use the collocations to validate an operational Ice Water Path (IWP) product from MHS measurements, produced by the National Environment Satellite, Data and Information System (NESDIS) in the Microwave Surface and Precipitation Products System (MSPPS). IWP values from the CloudSat CPR are found to be significantly larger than those from the MSPPS. Secondly, we compare the relation between IWP and MHS channel 5 (190.311 GHz) brightness temperature for two datasets: the collocated dataset, and an artificial dataset. We find a larger variability in the collocated dataset. Finally, we use the collocations to train an Artificial Neural Network and describe how we can use it to develop a new MHS-based IWP product. We also study the effect of adding measurements from the High Resolution Infrared Radiation Sounder (HIRS), channels 8 (11.11 μm) and 11 (8.33 μm). This shows a small improvement in the retrieval quality. The collocations described in the article are available for public use.

A reconfigurable neural environment on active networks

This paper proposes the deployment of a neural network computing environment on Active Networks. Active Networks are packet-switched computer networks in which packets can contain code fragments that are executed on the intermediate nodes. This feature allows the injection of small pieces of codes to deal with computer network problems directly into the network core, and the adoption of new computing techniques to solve networking problems. The goal of our project is the adoption of a distributed neural network for approaching tasks which are specific of the computer network environment. Dynamically reconfigurable neural networks are spread on an experimental wide area backbone of active nodes (ABone) to show the feasibility of the proposed approach.

Biogenic nitric oxide emissions from soils: impact on NOx and ozone over West Africa during AMMA (African Monsoon Multidisciplinary Analysis) - Modelling study

Nitrogen oxide biogenic emissions from soils are driven by soil and environmental parameters. The relationship between these parameters and NO fluxes is highly non linear. A new algorithm, based on a neural network calculation, is used to reproduce the NO biogenic emissions linked to precipitations in the Sahel on the 6 August 2006 during the AMMA campaign. This algorithm has been coupled in the surface scheme of a coupled chemistry dynamics model (MesoNH Chemistry) to estimate the impact of the NO emissions on NOx and O3 formation in the lower troposphere for this particular episode. Four different simulations on the same domain and at the same period are compared: one with anthropogenic emissions only, one with soil NO emissions from a static inventory, at low time and space resolution, one with NO emissions from neural network, and one with NO from neural network plus lightning NOx. The influence of NOx from lightning is limited to the upper troposphere. The NO emission from soils calculated with neural network responds to changes in soil moisture giving enhanced emissions over the wetted soil, as observed by aircraft measurements after the passing of a convective system. The subsequent enhancement of NOx and ozone is limited to the lowest layers of the atmosphere in modelling, whereas measurements show higher concentrations above 1000 m. The neural network algorithm, applied in the Sahel region for one particular day of the wet season, allows an immediate response of fluxes to environmental parameters, unlike static emission inventories. Stewart et al (2008) is a companion paper to this one which looks at NOx and ozone concentrations in the boundary layer as measured on a research aircraft, examines how they vary with respect to the soil moisture, as indicated by surface temperature anomalies, and deduces NOx fluxes. In this current paper the model-derived results are compared to the observations and calculated fluxes presented by Stewart et al (2008).

Benchmarking consensus model quality assessment for protein fold recognition

Background: Selecting the highest quality 3D model of a protein structure from a number of alternatives remains an important challenge in the field of structural bioinformatics. Many Model Quality Assessment Programs (MQAPs) have been developed which adopt various strategies in order to tackle this problem, ranging from the so called "true" MQAPs capable of producing a single energy score based on a single model, to methods which rely on structural comparisons of multiple models or additional information from meta-servers. However, it is clear that no current method can separate the highest accuracy models from the lowest consistently. In this paper, a number of the top performing MQAP methods are benchmarked in the context of the potential value that they add to protein fold recognition. Two novel methods are also described: ModSSEA, which based on the alignment of predicted secondary structure elements and ModFOLD which combines several true MQAP methods using an artificial neural network. Results: The ModSSEA method is found to be an effective model quality assessment program for ranking multiple models from many servers, however further accuracy can be gained by using the consensus approach of ModFOLD. The ModFOLD method is shown to significantly outperform the true MQAPs tested and is competitive with methods which make use of clustering or additional information from multiple servers. Several of the true MQAPs are also shown to add value to most individual fold recognition servers by improving model selection, when applied as a post filter in order to re-rank models. Conclusion: MQAPs should be benchmarked appropriately for the practical context in which they are intended to be used. Clustering based methods are the top performing MQAPs where many models are available from many servers; however, they often do not add value to individual fold recognition servers when limited models are available. Conversely, the true MQAP methods tested can often be used as effective post filters for re-ranking few models from individual fold recognition servers and further improvements can be achieved using a consensus of these methods.

Group stochastic search for object detection and tracking

A technique is presented for locating and tracking objects in cluttered environments. Agents are randomly distributed across the image, and subsequently grouped around targets. Each agent uses a weightless neural network and a histogram intersection technique to score its location. The system has been used to locate and track a head in 320x240 resolution video at up to 15fps.

Prediction of Parkinson's disease tremor onset using artificial neural networks

In this paper we present the initial results using an artificial neural network to predict the onset of Parkinson's Disease tremors in a human subject. Data for the network was obtained from implanted deep brain electrodes. A tuned artificial neural network was shown to be able to identify the pattern of the onset tremor from these real time recordings.

Identification of Parkinson’s disease tremor onset using artificial neural networks

In this paper we consider the possibility of using an artificial neural network to accurately identify the onset of Parkinson’s Disease tremors in human subjects. Data for the network is obtained by means of deep brain implantation in the human brain. Results presented have been obtained from a practical study (i.e. real not simulated data) but should be regarded as initial trials to be discussed further. It can be seen that a tuned artificial neural network can act as an extremely effective predictor in these circumstances.

Synapsing variable length crossover: An algorithm for crossing and comparing variable length genomes

The Synapsing Variable Length Crossover (SVLC) algorithm provides a biologically inspired method for performing meaningful crossover between variable length genomes. In addition to providing a rationale for variable length crossover it also provides a genotypic similarity metric for variable length genomes enabling standard niche formation techniques to be used with variable length genomes. Unlike other variable length crossover techniques which consider genomes to be rigid inflexible arrays and where some or all of the crossover points are randomly selected, the SVLC algorithm considers genomes to be flexible and chooses non-random crossover points based on the common parental sequence similarity. The SVLC Algorithm recurrently "glues" or synapses homogenous genetic sub-sequences together. This is done in such a way that common parental sequences are automatically preserved in the offspring with only the genetic differences being exchanged or removed, independent of the length of such differences. In a variable length test problem the SVLC algorithm is shown to outperform current variable length crossover techniques. The SVLC algorithm is also shown to work in a more realistic robot neural network controller evolution application.

Communicating neurons: a connectionist spiking neuron implementation of stochastic diffusion search

An information processing paradigm in the brain is proposed, instantiated in an artificial neural network using biologically motivated temporal encoding. The network will locate within the external world stimulus, the target memory, defined by a specific pattern of micro-features. The proposed network is robust and efficient. Akin in operation to the swarm intelligence paradigm, stochastic diffusion search, it will find the best-fit to the memory with linear time complexity. information multiplexing enables neurons to process knowledge as 'tokens' rather than 'types'. The network illustrates possible emergence of cognitive processing from low level interactions such as memory retrieval based on partial matching. (C) 2007 Elsevier B.V. All rights reserved.

A comparison between feature-based and EM-based contour tracking

Most active-contour methods are based either on maximizing the image contrast under the contour or on minimizing the sum of squared distances between contour and image 'features'. The Marginalized Likelihood Ratio (MLR) contour model uses a contrast-based measure of goodness-of-fit for the contour and thus falls into the first class. The point of departure from previous models consists in marginalizing this contrast measure over unmodelled shape variations. The MLR model naturally leads to the EM Contour algorithm, in which pose optimization is carried out by iterated least-squares, as in feature-based contour methods. The difference with respect to other feature-based algorithms is that the EM Contour algorithm minimizes squared distances from Bayes least-squares (marginalized) estimates of contour locations, rather than from 'strongest features' in the neighborhood of the contour. Within the framework of the MLR model, alternatives to the EM algorithm can also be derived: one of these alternatives is the empirical-information method. Tracking experiments demonstrate the robustness of pose estimates given by the MLR model, and support the theoretical expectation that the EM Contour algorithm is more robust than either feature-based methods or the empirical-information method. (c) 2005 Elsevier B.V. All rights reserved.

Prediction of Parkinson’s disease tremor onset using radial basis function neural networks

The possibility of using a radial basis function neural network (RBFNN) to accurately recognise and predict the onset of Parkinson’s disease tremors in human subjects is discussed in this paper. The data for training the RBFNN are obtained by means of deep brain electrodes implanted in a Parkinson disease patient’s brain. The effectiveness of a RBFNN is initially demonstrated by a real case study.