Adaptive Anomaly Detection for Internet of Things in Hierarchical Edge Computing: A Contextual-Bandit Approach

The advances in deep neural networks (DNN) have significantly enhanced real-time detection of anomalous data in IoT applications. However, the complexity-accuracy-delay dilemma persists: Complex DNN models offer higher accuracy, but typical IoT devices can barely afford the computation load, and the remedy of offloading the load to the cloud incurs long delay. In this article, we address this challenge by proposing an adaptive anomaly detection scheme with hierarchical edge computing (HEC). Specifically, we first construct multiple anomaly detection DNN models with increasing complexity and associate each of them to a corresponding HEC layer. Then, we design an adaptive model selection scheme that is formulated as a contextual-bandit problem and solved by using a reinforcement learning policy network . We also incorporate a parallelism policy training method to accelerate the training process by taking advantage of distributed models. We build an HEC testbed using real IoT devices and implement and evaluate our contextual-bandit approach with both univariate and multivariate IoT datasets. In comparison with both baseline and state-of-the-art schemes, our adaptive approach strikes the best accuracy-delay tradeoff on the univariate dataset and achieves the best accuracy and F1-score on the multivariate dataset with only negligibly longer delay than the best (but inflexible) scheme.

Applying Parallel and Distributed Models on Bio-Inspired Algorithms via a Clustering Method

In the world of optimization, especially concerning metaheuristics, solving complex problems represented by applying big data and constraint instances can be difficult. This is mainly due to the difficulty of implementing efficient solutions that can solve complex optimization problems in adequate time, which do exist in different industries. Big data has demonstrated its efficiency in solving different concerns in information management. In this paper, an approach based on multiprocessing is proposed wherein clusterization and parallelism are used together to improve the search process of metaheuristics when solving large instances of complex optimization problems, incorporating collaborative elements that enhance the quality of the solution. The proposal deals with machine learning algorithms to improve the segmentation of the search space. Particularly, two different clustering methods belonging to automatic learning techniques, are implemented on bio-inspired algorithms to smartly initialize their solution population, and then organize the resolution from the beginning of the search. The results show that this approach is competitive with other techniques in solving a large set of cases of a well-known NP-hard problem without incorporating too much additional complexity into the metaheuristic algorithms.

Real Time Fault Location Using the Retardation Method

A new method for short circuit fault location is proposed based on instantaneous signal measurement and its derivatives, and is based on the retardation phenomena. The difference between the times in which a signal is registered in two detectors is used to locate the fault. Although a description of faults in terms of a lumped circuit is useful for elucidating the methods for detecting the fault. This description will not suffice to describe the fault signal propagation hence a distributed models is needed which is given in terms of the telegraph equations. Those equations are used to derive a transmission line transfer function, and an exact analytical description of the short circuit signal propagating in the transmission line is obtained. The analytical solution was verified both by numerical simulations and experimentally.

Interpreting the Manning Roughness Coefficient in Overland Flow Simulations with Coupled Hydrological-Hydraulic Distributed Models

There is still little experience on the effect of the Manning roughness coefficient in coupled hydrological-hydraulic distributed models based on the solution of the Shallow Water Equations (SWE), where the Manning coefficient affects not only channel flow on the basin hydrographic network but also rainfall-runoff processes on the hillslopes. In this kind of model, roughness takes the role of the concentration time in classic conceptual or aggregated modelling methods, as is the case of the unit hydrograph method. Three different approaches were used to adjust the Manning roughness coefficient in order to fit the results with other methodologies or field observations—by comparing the resulting time of concentration with classic formulas, by comparing the runoff hydrographs obtained with aggregated models, and by comparing the runoff water volumes with observations. A wide dispersion of the roughness coefficients was observed to be generally much higher than the common values used in open channel flow hydraulics.

Rainfall spatial variability in the application of catchment morphing for ungauged catchments

Catchment Morphing (CM) is a newly proposed approach to apply fully distributed models for ungauged catchments and has been experimented in several catchments in the UK. As one of the most important input datasets for hydrological models, rainfall spatial variability is influential to the stream variabilities and simulation performance. A homogenous rainfall was utilized in the previous experiments with Catchment Morphing. This study applied a spatially distributed rainfall from CEH-GEAR rainfall dataset in the morphed catchment for ungauged catchments as the follow-on study. Three catchments in the UK were used for rainfall spatial analysis and CEH-GEAR rainfall data were adopted for additional spatial analysis. The results demonstrate the influence of rainfall spatial information to the model performance with CM and illustrate the ability of morphed catchment to tackle with spatially varied information. More spatially distributed information is expected to be introduced for a wider application of CM.

Estimation of Continuous Blood Pressure from PPG via a Federated Learning Approach

Ischemic heart disease is the highest cause of mortality globally each year. This puts a massive strain not only on the lives of those affected, but also on the public healthcare systems. To understand the dynamics of the healthy and unhealthy heart, doctors commonly use an electrocardiogram (ECG) and blood pressure (BP) readings. These methods are often quite invasive, particularly when continuous arterial blood pressure (ABP) readings are taken, and not to mention very costly. Using machine learning methods, we develop a framework capable of inferring ABP from a single optical photoplethysmogram (PPG) sensor alone. We train our framework across distributed models and data sources to mimic a large-scale distributed collaborative learning experiment that could be implemented across low-cost wearables. Our time-series-to-time-series generative adversarial network (T2TGAN) is capable of high-quality continuous ABP generation from a PPG signal with a mean error of 2.95 mmHg and a standard deviation of 19.33 mmHg when estimating mean arterial pressure on a previously unseen, noisy, independent dataset. To our knowledge, this framework is the first example of a GAN capable of continuous ABP generation from an input PPG signal that also uses a federated learning methodology.

Value-at-Risk Estimation of Equity Market Risk in India

The value-at-risk (Va) method in market risk management is becoming a benchmark for measuring “market risk” for any financial instrument. The present study aims at examining which VaR model best describes the risk arising out of the Indian equity market (Bombay Stock Exchange (BSE) Sensex). Using data from 2006 to 2015, the VaR figures associated with parametric (variance–covariance, Exponentially Weighted Moving Average, Generalized Autoregressive Conditional Heteroskedasticity) and non-parametric (historical simulation and Monte Carlo simulation) methods have been calculated. The study concludes that VaR models based on the assumption of normality underestimate the risk when returns are non-normally distributed. Models that capture fat-tailed behaviour of financial returns (historical simulation) are better able to capture the risk arising out of the financial instrument.

A physically-based distributed karst hydrological model (QMG model-V1.0) for flood simulations

Karst trough valleys are prone to flooding, primarily because of the unique hydrogeological features of karst landform, which are conducive to the spread of rapid runoff. Hydrological models that represent the complicated hydrological processes in karst regions are effective for predicting karst flooding, but their application has been hampered by their complex model structures and associated parameter set, especially so for distributed hydrological models, which require large amounts of hydrogeological data. Distributed hydrological models for predicting the Karst flooding is highly dependent on distributed structrues modeling, complicated boundary parameters setting, and tremendous hydrogeological data processing that is both time and computational power consuming. Proposed here is a distributed physically-based karst hydrological model, known as the QMG (Qingmuguan) model. The structural design of this model is relatively simple, and it is generally divided into surface and underground double-layered structures. The parameters that represent the structural functions of each layer have clear physical meanings, and the parameters are less than those of the current distributed models. This allows modeling in karst areas with only a small amount of necessary hydrogeological data. 18 flood processes across the karst underground river in the Qingmuguan karst trough valley are simulated by the QMG model, and the simulated values agree well with observations, for which the average value of Nash–Sutcliffe coefficient was 0.92. A sensitivity analysis shows that the infiltration coefficient, permeability coefficient, and rock porosity are the parameters that require the most attention in model calibration and optimization. The improved predictability of karst flooding by the proposed QMG model promotes a better mechanistic depicting of runoff generation and confluence in karst trough valleys.