Backscatter communications can enable the proliferation of Zero-energy (ZE) Internet of Things (IoT) services in 6G, ushering in a new era in green communications. To achieve a widescale adoption, backscatter communications needs to be integrated into the cellular ecosystem. This motivates the need to design a bistatic backscatter modulation technique which is not only compatible with the 3GPP New Radio (NR) physical layer but also robust to direct link interference. In this paper, a non-ambient bistatic backscatter communications system consisting of a cellular carrier emitter (CE), a backscatter device (BD) and a cellular reader (RD) is investigated. An analytical model is presented to explain the design rationale for an orthogonal frequency division multiplexing (OFDM) based physical layer for the considered system. Specifically, to achieve orthogonality between the CE and BD signals arriving concurrently at the RD, a comb-like OFDM subcarrier allocation is leveraged at the CE while a specially designed impedance switching sequence is applied at the BD to generate OFDM-compatible backscatter symbols with a binary or higher-order phase shift keying modulation. The link performance of the proposed system is evaluated in terms of the achievable bit error rate versus the signal-to-noise ratio at the RD for the CE-RD link and the BD-RD link under direct link interference. Simulation results confirm the viability of the proposed design for a cellular ZE-IoT system.

This work aims to disclose a closed-form analytical solution to attitude motion for a rigid body subject to zero body-fixed torques, i.e., Euler-Poinsot problem. Revisiting Routh's study, the presence of multiple solutions are identified. To verify the proposed solution, numerical simulations and real-life experiment results are presented. A Code Ocean repository is also provided so that readers could test the algorithm individually.

In this study, we provide the groundwork for using machine learning techniques in big software companies for evaluating and forecasting product quality.

People’s social acceptance and trust in robots are a direct consequence of people’s ability to infer and predict the robot’s behaviour.  However, there is no clear consensus on how the legibility of a robot’s behaviour and explanations should be assessed. In this work, the construct of Theory of Mind (i.e., the ability to attribute mental states to others) is taken into account, and a computerised version of the Theory of Mind Picture Sequencing Task is presented. Our tool, called the Human-Robot Interaction Video Sequencing Task (HRIVST), evaluates the legibility of a robot’s behaviour for humans by asking them to order short videos to form a logical sequence of the robot’s actions. To validate the proposed metrics, we recruited a sample of 86 healthy subjects. Results showed that the HRIVST has good psychometric properties and is a valuable tool for assessing the legibility of robot behaviours. We also evaluated the effects of symbolic explanations, the presence of a person during the interaction, and the humanoid appearance on a robot’s behaviour prediction accuracy. Results showed that the interaction condition had no effect on the legibility of the robot’s behaviour while the combination of humanoid robots and explanations displaying seems to result in better performances of the task.

The increasing popularity of ChatGPT and its widespread adoption across various industries have brought forth new and evolving challenges. In this research paper, our primary focus is to investigate how ChatGPT can benefit the Information Technology (IT) sector while exploring the challenges it presents. Given the significance of this topic, numerous Chief Information Security Officers (CISOs) are currently working on understanding and implementing ChatGPT. Thus, we embarked on this research endeavor to delve deeper into this subject. Our paper, titled “Unleashing the Potential: Overcoming Hurdles and Embracing Generative AI in IT Workplaces: Advantages, Guidelines, and Policies,” aims to examine the advantages, hurdles, and potential of Generative AI (GenAI) and its applications in IT environments. To accomplish our research goals, we employed multiple GenAI tools and conducted extensive investigations to explore their benefits and identify strategies to mitigate associated challenges. By leveraging these tools, we aimed to unlock the full potential of GenAI in the IT landscape. Our research paper delves into various aspects, including the advantages offered by GenAI, the hurdles encountered during its implementation, and the potential it holds for transforming IT workplaces. Furthermore, we provide guidelines and policies to ensure the responsible and effective utilization of GenAI within IT organizations. Through this research, we aim to contribute to the growing body of knowledge surrounding GenAI, enabling IT professionals to make informed decisions and harness the benefits of this emerging technology. By addressing the challenges and providing guidelines, we aim to facilitate the seamless integration of GenAI into IT workplaces, fostering innovation and efficiency in the industry. Please see attached PDF for result.

The no-insulation (NI) high temperature superconductor (HTS) winding technique enables the fabrication of highly compact magnets with self-quench protection. NI pancake coils are implemented to develop a 28 T HTS magnet for the operation of a 792 GHz gyrotron, which is a microwave source for dynamic nuclear polarization nuclear magnetic resonance. To this end, three NI all-GdBCO double pancake (DP) magnets were fabricated using different winding diameters and procedures. The objective was to explore different mechanical coil protection mechanisms at high magnetic fields, including clamping, overband, and solder impregnation. Experiments in liquid helium using magnets with winding diameters of 18 mm, 25 mm, and 66 mm yielded a center field of 14.4 T, 11.2 T, and 8.1 T, respectively. The maximum currents applied to the DP magnets ranged from 780 A to 1000 A. Both the 18 mm and 66 mm DP coils contained 400 m (2 × 200 m) HTS tape, while the 25 mm seamless DP coil consisted of only one continuous 200 m HTS tape. The 25 mm magnet with solder impregnation showed the best repeatability, although the current density was reduced owing to the solder thickness between the coil windings. Critical to the implementation of such coils in liquid helium is to effectively transfer high currents to the magnet, while not compromising the helium boil-off. Furthermore, the design of hybrid copper-HTS current leads capable of carrying current larger than 1000 A from room temperature to the HTS magnet at 4.2 K is presented. The implementation of liquid nitrogen-cooled hybrid copper-HTS current leads reduced the helium boil-off and permitted independent temperature control of the current leads.

Lithium-ion batteries (LIBs) immensely contribute to the electromobility’s success for achieving climate change goals. As LIBs are forecasted to succumb to optimization limits in the coming decade, next generation battery technologies, such as all-solid-state batteries (ASSBs), gain noteworthy attention for meeting ever-increasing cell performance requirements. By deploying solid electrolytes (SEs), compared to liquid electrolytes in current LIBs, ASSBs benefit from enhanced safety against flammability and allow for the usage of lithium metal anodes for higher energy densities. Here, polymer solid electrolytes, such Polyethylene oxide (PEO), are widely used for their high flexibility and hence beneficial processability properties compared other SEs. However, their adhesive behavior poses challenges when conducting handling and stacking processes with conventional grippers during cell assembly. In this research, we present a parameter study on ASSB handling and stacking with PEO-based cell components aiming to promote process understanding and point out optimization potentials. An experimental design for testing different grippers is devised by which deposition accuracy was systematically assessed in relation to the holding force, gripper speed, and placement distance. Within this evaluation, the electrostatic gripper with PTFE dielectric provides adequate position and orientation accuracies in almost all experiments while showing improved accuracies with higher holding forces. Parameter settings achieving higher overall deposition accuracies for all tested grippers are identified. This research provides insights into the establishment of stacking processes for realizing an industry-scale ASSB production.

This work explores the concept of a nonlinear extersion to the standard Maxwell equations for describing electrical and magnetic fields.  The nonlinear model has solutions of concentrated electromagnetic energy which would be characterized as particles in the model universe.  Some key findings are 1) negative and positive charge particles exist with the positive charge particle having a large energy compared to the negative charge particle 2) magnetic charge particles exist and are required to have a finite energy for the positive charge particle.  This publication is an initial study of the model equations with a general summary of the results.

In this paper, two crop datasets are investigated.  First one is numerical data which is downloaded crop dataset from   https://github.com/Shrey-B/AV-Janatahack-Machine-Learning-in-Agriculture obtained from US field data collection. The dataset contains 88858 labeled samples, eight features and three classes.  Second one is a collection of crop image data.  This dataset is downloaded from https://www.kaggle.com/datasets/aman2000jaiswal/agriculture .  Crop image dataset contains 1005 samples having five type of images namely maize, wheat, jute, rice and sugarcane. Each sample crop image consists of 224 × 224 pixels) of all category.

This work shows the simulation of a CMOS spiking neuron model that can mimic the behaviour of a biological neuron for both hyperpolarizing and depolarizing currents. The proposed schematic design can exhibit the post-inhibitory rebound excitation (PIRE) or anode break excitation (ABE) phenomena. The model has multiple control parameters for regulating the operation of a neuron. It is shown that the electrophysiological properties of a neuron, such as spiking frequency, spike height, resting potential, and mean inter-spike interval, can be varied by adjusting two control parameters of the developed schematic. Schematic of excitatory and inhibitory synapses to link two neuronal models are designed using CMOS. Average power dissipation and energy consumed per spike of the schematic designs are examined. Furthermore, PIRE is demonstrated for a strong inhibitory synaptic current. The significance of including PIRE in such a neuronal model is shown by its necessity in the generation of persistent activity in the network, which mimics the realization of working memory or attention. A network of such neuronal designs with the appropriate realization of the synaptic region has implications for human-to-machine interfaces and pattern recognition problems

This paper presents an evaluation of the mathematics performance of Google Bard on the Mathematics Test of the Vietnamese National High School Graduation Examination. The objective was to assess the accuracy and effectiveness of Google Bard in solving mathematical problems commonly found in the Vietnamese curriculum. To facilitate this evaluation, the VNHSGE Vietnamese dataset was translated into English since Bard does not currently support Vietnamese. The results indicate that Bard achieved a performance of 38.8%, lagging behind ChatGPT and BingChat, which achieved relatively higher scores of 58.8% and 60% respectively. It is important to note that this performance may be influenced by the translation process from Vietnamese to English. As a result, it is not recommended for Vietnamese students to rely on Google Bard for mathematics problem-solving.

One of the advantages of light controlling is the enhancement of its intensity in localization of the beam in the near- and far-fields. Our desire is to design a concentric nano elliptical apertures array tip (CNEAAT) for this purpose. The array tip manipulates the incident light by the nano elliptical apertures with sizes smaller than half-wavelength of the incident photons over the visible spectral range. The designed array tip generates the beam with the improved optical characteristics for maximum intensity enhancement at a special direction of far-field points. The light is steered to examine the output beam of the CNEAAT at various elevation angles in the near-field. For a special scanned impinged angle, the CNEAAT irradiation is focused on an assigned material under-test (MUT) for study of the beam enhancement. The MUT is inserted inside a cylindrical reservoir. The reservoir is inscribed by wall with n=1 and height of 40μm, includes the cross-section radius of 138nm and a metal bottom. In our design, the MUT is assumed to have real permittivity, as a sample to test the intensity enhancement of the impinged and transmission beams. The intensity enhancement obtained by the CNEAAT is anticipated comparable with the standardized apertures arrays. The design is found to be used in the numerous applications analogous to scanning near-field optical microscopy (SNOM) for imaging.

This paper presents TULIP, a new architecture for a variable precision Quantized Neural Network (QNN) inference. It is designed with the goal of maximizing energy efficiency per classification. TULIP is constructed by arranging a collection of unique processing elements (TULIP-PEs) in a single instruction multiple data (SIMD) fashion. Each TULIP-PE contains binary neurons that are interconnected using multiplexers. Each neuron also has a small dedicated local register connected to it. The binary neurons are implemented as standard cells and used for implementing threshold functions, i.e., an inner-product and thresholding operation on its binary inputs. The neurons can be reconfigured with a single change in the control signals to implement all the standard operations used in a QNN. This paper presents novel algorithms for implementing the operations of a QNN on the TULIP-PEs in the form of a schedule of threshold functions. TULIP was implemented as an ASIC in TSMC 40nm- LP technology. A QNN accelerator that employs a conventional MAC-based arithmetic processor was also implemented in the same technology to provide a fair comparison. The results show that TULIP is 30-50X more energy-efficient than an equivalent design, without any penalty in performance, area, or accuracy. Furthermore, TULIP achieves these improvements without using traditional techniques such as voltage scaling or approximate computing. Finally, the paper also demonstrates how the run- time trade-off between accuracy and energy efficiency is done on the TULIP architecture.

The practical design of Free-space optical communication (FSOs) faces significant challenges due to atmospheric turbulence and pointing errors. To overcome these obstacles, a hybrid technology combining FSO with NB-PLC scheme has been analyzed in this paper. The study evaluates the area coverage of a hybrid model of Power Line Communication (PLC) and Free Space Optical Communication (FSO) by measuring the Successful Transmission Probability (STP) and probability of error. The analysis assumes that the PLC channel undergoes Rayleigh fading, while FSO links are modeled using the Gamma?Gamma distribution. An algorithm is used to weigh each link by its Successful Transmission Probability (STP) and select the best possible path using the Single Channel Selection scheme (SCS). The Monte Carlo simulation results demonstrate that the pro?posed system is highly advantageous in smart city environments where channel fading is dynamic.

Providing high throughput and quality of service in modern stadiums necessitates the placement of hundreds of access points (APs). Optimizing the locations of APs in such venues via measurements requires significant resources. Even simulation methods, such as ray-tracing, can be computationally costly. We provide a solution to this problem by building a propagation model based on machine learning (ML) that rapidly predicts received signal strengths in stadiums. We train the model with a small set of simulated data generated by a ray-tracer. We use input features, such as the electrical distance between the transmitter and the receiver and the antenna gain along the direct path between the two, to generalize to new transmitter locations, antenna patterns and stadium geometries. Geometry and pattern generalization have not been included in existing propagation models for stadiums. Finally, we present a novel sampling approach for the input features in a given stadium, ensuring the computational efficiency and accuracy of the ML model. The results demonstrate the accuracy of our propagation model for new transmitter locations, patterns and stadiums. The trained model is also considerably faster than a ray-tracer, making it an efficient tool for resource planning tasks, such as optimal placement of APs.

This research paper aims to address the development of a cost-effective Data Platform tailored for analytics, data science, and AI applications. While the focus of this paper lies predominantly on the retail industry, the approach presented can be applicable across various domains. The advent of cloud technology has significantly benefited small and mid-sized organizations, leading to discussions surrounding IT modernization in recent years. Within the context of this study, we concentrate specifically on small and mid-sized retail organizations that utilize ERP, allocation, Warehouse Management, and Sell applications, both in-store and online. The Information Technology (IT) sector strives to establish a scalable data platform that caters to reporting, data science, and AI initiatives. This research paper places special emphasis on mid-sized organizations, which often face budget constraints and stringent project timelines. Establishing an IT organization with limited resources and budget necessitates careful consideration of the available skill set. By addressing the challenges and opportunities associated with building a cost-effective Data Platform, this research paper aims to provide valuable insights and practical guidance for organizations operating within similar contexts. The subsequent sections will delve into the methodologies, technologies, and strategies employed to achieve a scalable and efficient data platform that aligns with the specific requirements of small and mid-sized retail organizations. More detail will be found at attached PDF.

Machine Learning (ML) has shown its capacity for solving problems where prediction is a key factor, and the fact led us to start to explore ML in Data-prefetching, where prediction is also a key factor. Among various ML algorithms, Perceptron Learning caught our attention because of its simplicity that could make it fit well for microarchitecture design. Therefore, we propose Perceptron based Data-prefetching in this paper, contributing to the state of the art by presenting the followings: a novel memory address prediction algorithm based on Perceptron Learning; the hardware implementation thereof; and the evaluation results with an industry-standardized benchmark suite.

This paper discusses the rapid advancement of artificial intelligence (AI) in the form of generative pre-trained transformer (GPT) technology such as the latest GPT-4 platform. It highlights the potential of GPT technology to drastically change aspects of society, including creativity, problem-solving, employment, education, justice, medicine, and governance. The author emphasizes the need for policymakers and experts to join in regulating against the potential risks and implications of this technology. The European Commission has taken steps to address the risks of AI through the European AI Act (EIA), which categorizes AI uses based on their potential harm. The legislation aims to ensure scrutiny and control in extreme cases like autonomous weapons or medical devices. However, the author criticizes the lack of meaningful AI oversight in the United States and argues that time has come for government to step in if it desires to make any meaningful change given the technology’s (1) rate of diffusion, (2) multiple products GPT technology is anticipated to augment, and (3) depths to which it may penetrate daily life, including countless fields of employment.