Hardware-accelerated Simulation-based Inference of Stochastic Epidemiology Models for COVID-19

Epidemiology models are central to understanding and controlling large-scale pandemics. Several epidemiology models require simulation-based inference such as Approximate Bayesian Computation (ABC) to fit their parameters to observations. ABC inference is highly amenable to efficient hardware acceleration. In this work, we develop parallel ABC inference of a stochastic epidemiology model for COVID-19. The statistical inference framework is implemented and compared on Intel’s Xeon CPU, NVIDIA’s Tesla V100 GPU, Google’s V2 Tensor Processing Unit (TPU), and the Graphcore’s Mk1 Intelligence Processing Unit (IPU), and the results are discussed in the context of their computational architectures. Results show that TPUs are 3×, GPUs are 4×, and IPUs are 30× faster than Xeon CPUs. Extensive performance analysis indicates that the difference between IPU and GPU can be attributed to higher communication bandwidth, closeness of memory to compute, and higher compute power in the IPU. The proposed framework scales across 16 IPUs, with scaling overhead not exceeding 8% for the experiments performed. We present an example of our framework in practice, performing inference on the epidemiology model across three countries and giving a brief overview of the results.

Towards Named Data Networking for Internet of Things Search Engines

The advance of internet of things (IoT) techniques enables a variety of smart-world systems in energy, transportation, home, and city infrastructure, among others. To provide cost-effective data-oriented service, internet of things search engines (IoTSE) have received growing attention as a platform to support efficient data analytics. There are a number of challenges in designing efficient and intelligent IoTSE. In this chapter, the authors focus on the efficiency issue of IoTSE and design the named data networking (NDN)-based approach for IoTSE. To be specific, they first design a simple simulation environment to compare the IP-based network's performance against named data networking (NDN). They then create four scenarios tailored to study the approach's resilience to address network issues and scalability with the growing number of queries in IoTSE. They implement the four scenarios using ns-3 and carry out extensive performance evaluation to determine the efficacy of the approach concerning network resilience and scalability. They also discuss some remaining issues that need further research.

Achieve Efficient and Privacy-Preserving Compound Substring Query over Cloud

The cloud computing technique, which was initially used to mitigate the explosive growth of data, has been required to take both data privacy and users’ query functionality into consideration. Searchable symmetric encryption (SSE) is a popular solution that can support efficient attribute queries over encrypted datasets in the cloud. In particular, some SSE schemes focus on the substring query, which deals with the situation that the user only remembers the substring of the queried attribute. However, all of them just consider substring queries on a single attribute, which cannot be used to achieve compound substring queries on multiple attributes. This paper aims to address this issue by proposing an efficient and privacy-preserving SSE scheme supporting compound substring queries. In specific, we first employ the position heap technique to design a novel tree-based index to support substring queries on a single attribute and employ pseudorandom function (PRF) and fully homomorphic encryption (FHE) techniques to protect its privacy. Then, based on the homomorphism of FHE, we design a filter algorithm to calculate the intersection of search results for different attributes, which can be used to support compound substring queries on multiple attributes. Detailed security analysis shows that our proposed scheme is privacy-preserving. In addition, extensive performance evaluations are also conducted, and the results demonstrate the efficiency of our proposed scheme.

Quantitative prediction of conditional vulnerabilities in regulatory and metabolic networks using PRIME

The ability of Mycobacterium tuberculosis (Mtb) to adopt heterogeneous physiological states underlies its success in evading the immune system and tolerating antibiotic killing. Drug tolerant phenotypes are a major reason why the tuberculosis (TB) mortality rate is so high, with over 1.8 million deaths annually. To develop new TB therapeutics that better treat the infection (faster and more completely), a systems-level approach is needed to reveal the complexity of network-based adaptations of Mtb. Here, we report a new predictive model called PRIME (Phenotype of Regulatory influences Integrated with Metabolism and Environment) to uncover environment-specific vulnerabilities within the regulatory and metabolic networks of Mtb. Through extensive performance evaluations using genome-wide fitness screens, we demonstrate that PRIME makes mechanistically accurate predictions of context-specific vulnerabilities within the integrated regulatory and metabolic networks of Mtb, accurately rank-ordering targets for potentiating treatment with frontline drugs.

A new data classification improvement approach based on kernel clustering

Data classification is one of the most critical issues in data mining with a large number of real-life applications. In many practical classification issues, there are various forms of anomalies in the real dataset. For example, the training set contains outliers, often enough to confuse the classifier and reduce its ability to learn from the data. In this paper, we propose a new data classification improvement approach based on kernel clustering. The proposed method can improve the classification performance by optimizing the training set. We first use the existing kernel clustering method to cluster the training set and optimize it based on the similarity between the training samples in each class and the corresponding class center. Then, the optimized reliable training set is trained to the standard classifier in the kernel space to classify each query sample. Extensive performance analysis shows that the proposed method achieves high performance, thus improving the classifier’s effectiveness.

DeepFMD: Computational Analysis for Malaria Detection in Blood-Smear Images Using Deep-Learning Features

Malaria is one of the most infectious diseases in the world, particularly in developing continents such as Africa and Asia. Due to the high number of cases and lack of sufficient diagnostic facilities and experienced medical personnel, there is a need for advanced diagnostic procedures to complement existing methods. For this reason, this study proposes the use of machine-learning models to detect the malaria parasite in blood-smear images. Six different features—VGG16, VGG19, ResNet50, ResNet101, DenseNet121, and DenseNet201 models—were extracted. Then Decision Tree, Support Vector Machine, Naïve Bayes, and K-Nearest Neighbour classifiers were trained using these six features. Extensive performance analysis is presented in terms of precision, recall, f-1score, accuracy, and computational time. The results showed that automating the process can effectively detect the malaria parasite in blood samples with an accuracy of over 94% with less complexity than the previous approaches found in the literature.

Performance Characterization of Load Balancing Algorithm

The developing interest in mobile services increases the demand for well-planned and cautiously managed wireless local area networks (WLAN) deployment. In WLAN, a station can access services of the network through an access point (AP) after associating with it. Any number of access points can be accessed by the station whose signal strength is available from among the APs. But practically, a WLAN station (STA) always associates with the access point with higher signal strength among the APs. In WLAN, mobile stations continuously change their location, which results in an uneven network load allocation. This uneven load dissemination prompts an extensive performance degradation of WLAN. This paper presents mathematical modeling to characterize the WLAN performance by balancing the network load and enhancing network throughput. Riverbed Modeler simulator was used to investigate the performance parameters as network load and throughput of the network.

Implementation of facial emotion recognition using cnn on jetson TX2

In this paper, a convolutional neural network (CNN), one of the most popular deep learning architectures used for facial extraction research, has been implemented on NVIDIA Jetson TX2 hardware. Different from many existing approaches investigating CNN with complex structure and large parameters, we have focused on building a robust neural network through extensive performance comparison and evaluation. In addition, we have collected a dataset using a built-in camera on a laptop computer. Specifically, we have applied our model on Jetson TX2 hardware to take advantage of the computational power of the embedded GPU to optimize computation time and data training. In particular, both FER2013 and RAF datasets with seven basic emotions have been used for training and testing purposes. Finally, the evaluation results show that the proposed method achieves an accuracy of up to 72% on the testing dataset.

Towards Scalable and Efficient Architecture for Modeling Trust in IoT Environments

The Internet of Services (IoS) is gaining ground where cloud environments are utilized to create, subscribe, publish, and share services. The fast and significant evolution of IoS is affecting various aspects in people’s life and is enabling a wide spectrum of services and applications ranging from smart e-health, smart homes, to smart surveillance. Building trusted IoT environments is of great importance to achieve the full benefits of IoS. In addition, building trusted IoT environments mitigates unrecoverable and unexpected damages in order to create reliable, efficient, stable, and flexible smart IoS-driven systems. Therefore, ensuring trust will provide the confidence and belief that IoT devices and consequently IoS behave as expected. Before hosting trust models, suitable architecture for Fog computing is needed to provide scalability, fast data access, simple and efficient intra-communication, load balancing, decentralization, and availability. In this article, we propose scalable and efficient Chord-based horizontal architecture. We also show how trust modeling can be mapped to our proposed architecture. Extensive performance evaluation experiments have been conducted to evaluate the performance and the feasibility and also to verify the behavior of our proposed architecture.

Adaptive Waveform Design with Multipath Exploitation Radar in Heterogeneous Environments

The problem of detecting point like targets over a glistening surface is investigated in this manuscript, and the design of an optimal waveform through a two-step process for a multipath exploitation radar is proposed. In the first step, a non-adaptive waveform is transmitted anda constrained Generalized Likelihood Ratio Test (GLRT) detector is deduced at reception which exploits multipath returns in the range cell under test by modelling the target echo as a superposition of the direct plus the multipath returns. Under the hypothesis of heterogeneous environments, thus by assuming a compound-Gaussian distribution for the clutter return, this latter is estimated in the range cell under test through the secondary data, which are collected from the out-of-bin cells. The Fixed Point Estimate (FPE) algorithm is applied in the clutter estimation, then used to design the adaptive waveform for transmission in the second step of the algorithm, in order to suppress the clutter coming from the adjacent cells. The proposed GLRT is also used at the end of the second transmission for the final decision. Extensive performance evaluation of the proposed detector and adaptive waveform for various multipath scenarios is presented. The performance analysis prove that the proposed method improves the Signal-to-Clutter Ratio (SCR) of the received signal, and the detection performance with multipath exploitation.