We show for the first time that when a linearly polarized laser light incidents either a grating or a double slit, the output light is non-polarized. Namely, both grating and double slit depolarize the linearly polarized laser light.

Neural networks have been commonly used in different domains. Hardware implementation of the neural networks is challenging to fit different applications’ requirements. This often necessitates realizing different neural network FPGA configurations from the very scratch specific to a given application. This paper proposes a flexible method that can be reconfigured in a self-supervised way to fit several ap?plications’ requirements, by providing only maximum available computational nodes a prior. The proposed method is based on reconfiguring the required number of hidden layers and hidden nodes of that layer based on a given application. Reconfigurability also allows to decide and update the number of active nodes, at layer (l+1), which receives computational results of l th layer nodes. A configuration block is used to send a signal to the control block which follows each node output. The control block decides which following nodes receive the output according to the corresponding signal value from the configuration block. Each node receives a signal to be enabled or disabled from the configuration block. Therefore, the number of inputs of each node and nodes in each layer is determined according to the desired performance. The goal here is to automatically propose the optimal NN configuration through reconfigurability to fit different applications, to achieve the maximum accuracy possible. The optimality can be demonstrated in terms of minimum average power, average delay, and area overhead as well as maximum throughput and accuracy. The main advantage of the proposed approach is the reusability of the optimized architecture (which is self-learned online) for the very first application/dataset, as the initial configuration for any new/next dataset/application rather than starting from scratch. We have demonstrated with experimental results, the proposed approach helps primarily to significantly reduce optimized architecture search cost (the number of online training iterations) as well as associated average power consumption for successive datasets/applications. Our proposed method demonstrates its effectiveness both quantitatively and qualitatively. The proposed method is verified against two different classification problems, MNIST and CIFAR-10. Our proposed reconfigurable method demonstrates stable accuracy of 98.97% and 98.95% against the state-of-the-art neural network with two different fixed configurations as (98.85% and 73.0%) for MNIST and (93.47% and 70.21%) for CIFAR-10, respectively. The proposed method also demonstrates a 20.9% reduction in average power dissipation against the state-of-the-art method. The proposed method is implemented and tested using VHDL and Altera FPGA. The results show resource utilization is comparable with the state-of-the-art method.

We demonstrate that all enterprises of embodied cognition, ranging from the ‘simply’ cognitive to the conscious, and from the individual to the mechanical, institutional, and composite, are susceptible to induced patterns of hallucination, panic, and exhaustion on challenging landscapes of fog, friction, and deadly adversarial intent. We further argue, in consonance with C.S. Gray, that dreams of ‘Wunderwaffen’, ‘Wundertaktik’, and implacable ‘Reflexive Control’ are themselves siren hallucinations.

For local energy markets to become a widely accepted, real-life phenomenon, the models must extend beyond the techno-economical aspects and energy justice within these market structures must be thoroughly addressed. To measure the perceived fairness of a local energy market, some studies have adapted fairness indicators originated from other fields. However, it is still unclear how these indicators can accurately reflect distributional fairness in various local energy trading scenarios. This study conducts a comprehensive review and qualitative assessment of the existing fairness indicators in the local energy market literature and their ability to accurately capture different situations in the market. The study results in a discussion on their strength and limitations, and an encouragement to alter the indicators to better quantify energy justice in a local energy market context.

In this work, we have proposed bit-interleave SRAM-based IMC architecture for AI edge devices.

The preprint presents a novel sensing approach to time-resolved fluorescence imaging. It is also supported by supplementary materials.

The ShuttlePump is a novel implantable total artificial heart (TAH) concept based on a Linear-Rotary Actuator (LiRA) and currently under development at the Power Electronic Systems Laboratory, ETH Zurich in close partnership with Charit´e Berlin and the Medical University of Vienna. This paper presents the analysis, design and realization of the ShuttlePump Linear Actuator (LA) part, which is necessary to provide about 45 N of axial actuation force. Design criteria are minimization of volume and generated power losses in the winding, which could result in excess heating and/or blood damage, i.e. protein denaturation and aggregation. The LA is implemented as a Tubular LA (TLA) to maximize the active area for linear/axial force generation. After a preliminary analysis based on first principles, the TLA is optimized in detail with the aid of FEM simulations. The experimental measurements conducted on the realized TLA prototype verify the FEM simulation results and confirm the suitability for the realization of the ShuttlePump TAH.

Agent-based modeling is a powerful computational tool that can simulate complex behavior based on rules in both micro and macro scales. Defining the constant agent behavior patterns rules is uncertain since it depends on many facts leading to complicated decision-making processes. Integrating machine learning, specifically into a complex multiscale multi-agent model, could provide scientists with practical tools to make decisions. It also allows him to execute unlimited datasets in a large parametric ABM simulation on HPC/ Cloud infrastructures to improve accuracy, accelerate time to solution and significantly reduce costs. This paper presents a modeling and simulation systems workflow that allows us to mine network measures and visualizing the evolving behavior of a tumor agent based model at both molecular and cellular levels. This workflow is proposed to construct a complex, scalable network based on our simulated tumor growth agent-based model. Three applicable machine learning techniques are  combined with the traditional simulation to train the tumor network evolving. These techniques also discover and classify subgraphs. In this Model, the front-end integration is the strength of Python modeling based on packages such as an ABM package (Mesa), analysis, and visualization packages (Numpy, SciPy, Matplotlib). On the other hand, the Tumor simulation is getting more orchestrated while united with machine learning packages (NetworkX) or distribution utility modules and third-parties (Pypi) and container facilities deep at the back-end. Link to the code : https://github.com/ghta1000/Mesa-NetworkX-Agent-model.git

This letter introduces an attention-based unsupervised generative adversarial network, named Illu-GAN, for low- light camera image enhancement. Many images captured by camera sensors suffer from inadequate lighting conditions or over-/under- exposure. Previous low-light enhancement methods are mostly based on supervised learning-based and heavily rely on paired data. Existing unsupervised methods only utilize information in spatial domain and take no account of frequency information, which leads to results of poor quality. In this letter, we propose a new network, Illu- GAN, to incorporate the features from frequency and spatial domains to better guide the network through a Wavelet Transform module. Extensive experiments with Illu-GAN on various benchmarking datasets demonstrate qualitatively and quantitatively the improvement with respect to state-of-the-art methods. The qualitative and quantitative results demonstrate that the newly proposed Illu-GAN can generate more natural enhanced images, less noisy, and with better generalisation ability.

Compressed sensing (CS)-based techniques have been widely applied in the grant-free non-orthogonal multiple access (NOMA) to a single-antenna base station (BS). In this paper, we consider the multi-antenna reception at the BS for uplink grant-free access for the massive machine type communication (mMTC) with limited channel resources. To enhance the overloading performance of the BS, we develop a general framework for the synergistic amalgamation of the spatial division multiple access (SDMA) technique with the CS-based grant-free NOMA. We derive a closed-form statistical beamforming and a dynamic beamforming scheme for the inter-cluster interference suppression when applying SDMA. Based on this, we further develop a joint adaptive beamforming and subspace pursuit (J-ABF-SP) algorithm for the multiuser detection and data recovery, with a novel sparsity level decision method without the accurate knowledge of the noise level. To further improve the data recovery performance, we propose an interference cancellation-based J-ABF-SP scheme (J-ABF-SP-IC) by using the initial signal estimates generated from the J-ABF-SP algorithm. Extensive simulation results verify the superior user detection and signal recovery performance of our proposed algorithms in comparison with existing CS-based grant-free NOMA techniques.

Modern mobile communication systems exploit complicated modulations to increase spectrum efficiency and larger antenna arrays to suppress co-channel interferences. However, these techniques have also caused surges of power consumption and hardware cost, making it unsustainable to squeeze gains from the radio-frequency electronics. To circumvent the dilemma, optical mobile communication (OMC) systems exploit wideband lasers to carry data, and do not necessarily involve high-order modulation and oversized arrays. In OMC systems, parallel laser beams can achieve data rate of Tbps, connection density of thousands of links/m$^2$, and energy efficiency of Gbits/J, with low cost and low signal processing complexity, as long as the challenges in receiver tracking, channel decorrelation and anti-blocking can be tackled. To this end, this work presents a new OMC system architecture and discusses enabling technologies of practical OMC systems. A prototype is built to verify the feasibility of several key concepts. It is shown that a practical OMC system requires collaboration of fast and accurate positioning systems, rapid laser steerers, optical reflective intelligent surfaces and decorrelation optics.

In the present scenario, Women journalists have played a significant role in shaping the media industry, from reporting on important events to advocating for gender equality and social justice. However, despite progress in recent years, women journalists still need to overcome several challenges and threats in the industry. Gender discrimination is one major challenge faced by women journalists today. This can take many forms, from being paid less than their male counterparts to being passed over for promotion. Women journalists may also face bullying and harassment in the workplace, which can make it difficult to do their job effectively. Another challenge women journalist's face is the underrepresentation of women in decision-making roles. Women are often excluded from leadership positions in media organizations, making it difficult for them to have their voices heard and advance in their careers. Women journalists have made significant contributions to the media industry and continue to do so. They bring a unique perspective to news coverage and have helped to create a more diverse and inclusive media landscape. As the industry continues to evolve, ensuring women have equal opportunities to succeed and thrive as journalists are important. This review paper highlights the opportunities and challenges for women journalists in the media industry.

As deep neural networks (DNNs) are being applied to a wide range of edge intelligent applications, it is critical for edge inference platforms to have both high-throughput and low-latency at the same time. Such edge platforms with multiple DNN models pose new challenges for scheduler designs. First, each request may have different service level objectives (SLOs) to improve quality of service (QoS). Second, the edge platforms should be able to efficiently schedule multiple heterogeneous DNN models so that system utilization can be improved. To meet these two goals, this paper proposes BCEdge, a novel learning-based scheduling framework that takes adaptive batching and concurrent execution of DNN inference services on edge platforms. We define a utility function to evaluate the trade-off between throughput and latency. The scheduler in BCEdge leverages maximum entropy-based deep reinforcement learning (DRL) to maximize utility by 1) co-optimizing batch size and 2) the number of concurrent models automatically. Our prototype implemented on different edge platforms shows that the proposed BCEdge enhances utility by up to 37.6% on average, compared to state-of-the-art solutions, while satisfying SLOs.

LIDO is a Metamodeling language for developing IoT solutions. It is a set of concepts of high-level abstraction. LIDO make it possible that people with different technical backgrounds and IoT culture can come together around the development of IoT solutions. LIDO allow people of different disciplines, having different viewpoints to come together to discuss and reflect on the modeling and implementation of an IoT solution. This article presents the philosophy underlying LIDO that makes it a universal language allowing sustainable development.

We develop a deep neural network (DNN) modeling methodology to predict the radiated emissions of a shielding enclosure in terms of its aperture attributes including aperture shape, size, pitch, and quantity. The target structure is the inside of a three-dimensional (3D) enclosure comprised of perfect electric conductor (PEC) boundaries with dimensions of a desktop personal computer (PC) containing thermal dissipation apertures on the surface of its back panel. The DNN model is developed to compute the radar cross section (RCS) as a function of aperture attributes and to enable the efficient assessment of the PC’s electromagnetic compatibility (EMC). To generate training data for machine learning (ML), we implement the modified equivalent current approximation (MECA) method and validate it against analytical methods and a commercial field-solver. We use MECA to compute RCS data for approximately 55,000 experiments across a wide range of aperture attributes. We examine numerous DNN models across parameters such as number of layers and nodes per layer, activation function, optimization algorithm, loss function, batch size, and epoch, to identify the optimal DNN model based on (a) accuracy, (b) computation time, and (c) memory usage. Results show excellent agreement between MECA and DNN predictions for previously unseen cases.

The residential services in the context of smart cities accu- mulate massive real-time inquiry data in natural language to describe the services in need. Such inquiry requests have diverse topics in the content and considerable variate length. Besides, the responsible departments that may handle the inquiry involve a large number of organizations, from metropolitan administration to local communities. Hence accumu- lated request data is primary and central to the service of accurately dispatching requests to responsible departments. The challenge is de- vising a data centric approach to fit the data with SOTA models and improve the request classification accuracy. In this paper, we analyze the factors of embedding tokens, data segmentation, model structures, and classification methods. We devise a unified modelling process with mul- tiple dataflows that combine these factors to observe their interactions. The experiment results demonstrate the compound effects and provide insights into how SOTA models respond differently to variations in these factors. The observations allow us to fine-tune the learning task at each stage and achieve a maximum 82.4% F1-Score.

Millimeter-wave (mmWave) beamspace massive multiple-input multiple-output (mMIMO) system with lens antenna array can minimize transceiver hardware complexity without compromising performance. However, the number of supported portable user terminals (PUTs) cannot exceed the number of radio frequency (RF) blocks accessible at the same time, frequency, and coding resources. As a result, we propose the integration of rate-splitting multiple access (RSMA) into a beamspace mMIMO system to support a larger number of PUTs than the number of available RF blocks while minimizing intra-beam interferences. Moreover, orthogonal random precoding (ORP) is utilized in the downlink of the beamspace mMIMO system to mitigate the inter-beam interferences and extend the cell coverage area. Then, we develop an optimization problem to optimize the system’s overall throughput while keeping the minimum needed throughput and power budget in consideration. The nonconvex optimization issue is then turned into a convex optimization problem using the successive convex approximation approach. Following that, we offer an alternating method to solve the approximate optimization issue and select an optimal solution. Furthermore, the suggested method’s effectiveness is evaluated in terms of total throughput, energy efficiency, and cell coverage area. Finally, numerical results confirm the superior performance of the proposed method over benchmark techniques in terms of sum throughput, energy efficiency, and cell coverage area.

We  consider active hypothesis testing with multiple heterogeneous agents. Each agent has access to its own set of experiments, has different action costs and forms its own beliefs. Additionally,  each experiment carries a global cost, and the agents must try to keep the expected cumulative cost below a certain threshold. We study a centralized and a decentralized scenario. Under the centralized scenario, the agents  are instructed how to act by a central controller. Under the decentralized scenario they communicate and exchange information over a directed graph. For each scenario we propose  separate deep reinforcement learning algorithms based on proximal policy optimization. Solutions to the decentralized problem start from fully decentralised training, and progressively introduce two levels of centralisation during training.  We assess the proposed algorithms in an example of anomaly detection over sensor networks, considering three different decentralised communication settings. We infer that all algorithms achieve the required accuracy level considerably faster than a single deep reinforcement learning agent, while satisfying the expected cost constraints when required. Under the assumption that the communication  graph is fully connected, the decentralised agents perform just as well, and sometimes better than the centralised  controller. Out of all decentralised algorithms, the one that uses a global critic is by far the best performing and can compete with the central controller even when the communication graph is not complete.