In today's digitized world, people rely heavily on numerous smart machines to perform everyday tasks. The number of smart devices has surged recently, leading to an increase in security vulnerabilities. Among these, the "Low-rate denial of service (LDoS)" attack stands out as particularly dangerous due to its stealthy and varied nature, posing significant challenges for current intrusion detection systems. This research introduces a hybrid approach to investigate LDoS attack features, combining hyperparameter optimization (HPO) with principal component analysis (PCA). To address dataset imbalance, the SMOTE technique is applied. PCA is used for dimensionality reduction, with the key hyperparameter 'n_components' optimized through HPO. The study utilizes the 'CICIDS2017' dataset, highlighting the importance of dimension reduction for improved performance. The hybrid method, termed HPO-S-PCA, is employed to analyze LDoS traffic features and extract relevant features. Using these extracted features, machine learning classifiers such as 'Logistic Regression (LR)', 'Support Vector Machine (SVM)', 'Decision Tree (DT)', 'Random Forest (RF)', 'K-Nearest Neighbors (KNN)', 'Kernel SVM', and 'Naive Bayes (NB)' were trained to detect LDoS attacks. Among these, SVM and KNN classifiers achieved more than 99.5% detection rate for positive anomalies. K-Nearest Neighbors outperforms all. The research observed a trade-off between True Positive Rate (TPR) and accuracy in existing studies and focused on enhancing both performance metrics through the novel hybrid approach.

The growing industry and its complex and large information sets require Big Data (BD) technology and its open-source frameworks (Apache Hadoop) to (1) collect, (2) analyze, and (3) process the information. This information usually ranges in size from gigabytes to petabytes of data. However, processing this data involves web consoles and communication channels which are prone to intrusion from hackers. To resolve this issue, a novel machine learning (ML)-based security-centric approach has been proposed to evade cyber-attacks on the Hadoop ecosystem while considering the complexity of Big Data in Cloud (BDC). An Apache Hadoop-based management interface "Ambari" was implemented to address the variation and distinguish between attacks and activities. The analyzed experimental results show that the proposed scheme effectively (1) blocked the interface communication and retrieved the performance measured data from (2) the Ambari-based virtual machine (VM) and (3) BDC hypervisor. Moreover, the proposed architecture was able to provide a reduction in false alarms as well as cyber-attack detection.

Ferroelectric Superconducting Quantum Interference Device (Fe-SQUID) has recently emerged as a viable option to realize superconducting computing due to its voltage-controlled switching, which is essential to build large-scale digital circuits. This is the first work to model Fe-SQUID-based logic circuits and develop standard cell libraries compatible with existing EDA tool flows. We provide a comprehensive evaluation of the power consumption and performance of a wide range of Fe-SQUIDbased circuits, benchmarking them against the state-of-the-art 5 nm FinFET-based circuits. Our 5 nm FinFET transistor model is validated against industrial measurements. The validation is conducted not only at room temperature but also at extremely low temperatures, down to 10K, for fair comparisons against Fe-SQUID superconducting circuits. Our findings revealed that contrary to CMOS-based circuits, circuits realized using Fe-SQUID dissipate significantly more power. This presents a substantial challenge within the constraints of limited cooling power budgets in state-of-the-art cryostats. Open source: we will release our developed Fe-SQUID compact model, Fe-SQUID standard cell libraries, and Fe-SQUID-based circuits and make them publicly available in a github repository.

Civilian communication during disasters such as earthquakes, floods, and military conflicts is crucial for saving lives. Nevertheless, several challenges exist during these circumstances such as the destruction of cellular communication and electricity infrastructure, lack of line of sight (LoS), and difficulty of localization under the rubble. In this article, we discuss key enablers that can boost communication during disasters, namely, satellite and aerial platforms, redundancy, silencing, and sustainable networks aided with wireless energy transfer (WET). The article also highlights how these solutions can be implemented in order to solve the failure of communication during disasters. Finally, it sheds light on unresolved challenges, as well as future research directions.

Facial expression analysis has been shown to be an unintrusive and cheap method of player experience modelling. Output from such systems is then often closely tied to measures of basic discrete emotions. Despite this, little games user research attempts to predict the intensity of players' emotions collected in response to validated measures of discrete emotion, that can be more widely interpreted than those used in game specific settings. Prediction of experience as a function of discrete emotions using player facial expressions therefore presents opportunities for both the furthering of intelligent game adaptation, and the fields of serious games and emotional regulation. This study tests multiple approaches for the use, and selection, of features describing players' gameplay and facial activity during an online multiplayer game. Over 18 hours of visual and behavioural data collected from 37 participants is used to build predictors of the intensity for discrete emotions of relaxation, happiness, and desire as measured using the Discrete Emotions Questionnaire, reaching a model accuracy of 55%, 57% and 63% respectively. Models using facial expression features collected over the course of entire game levels show significantly better accuracy than models using gameplay or facial reaction features.

With the evolution of 6G networks, the convergence of the Metaverse and video streaming has opened new avenues for immersive user experiences. However, delivering high-quality video streaming in such an interactive and virtual environment poses significant challenges, particularly with regards to Quality of Experience (QoE). This paper investigates the impact of user interaction behaviors on video streaming performance in the Metaverse, specifically over 6G networks. Through a comprehensive analysis of the factors affecting QoE-such as latency, bandwidth, and interactive response times-this study proposes a novel QoE optimization framework. The framework integrates user interaction data with network resource allocation strategies, leveraging the capabilities of 6G, including ultralow latency and high-speed communication. Experimental results demonstrate improved QoE for Metaverse video streaming when user interaction behaviors are taken into account, providing insights for future developments in immersive media over 6G.

Snow over sea ice affects the overall radiative balance of the polar region by modulating the planetary albedo and influencing sea ice growth. In satellite altimetry, snow is a major source of uncertainty in sea ice thickness estimates, especially with Ku-band radar altimeters. This is because conventional algorithms assume the primary scattering occurs at the snow/ice interface which may not hold for all snowpacks. In particular, complex snow morphological features, including wet snow and/or ice lenses, can shift the main scattering horizon toward the snow surface. These snow features may be particularly prominent over Antarctic sea ice. This study investigates how the properties of the snowpack in the Weddell Sea influence Ku-and Ka-band altimeter waveforms, particularly in the context of extreme conditions and complex snow stratigraphy. Using an adapted waveform model, we deconvolute radar echoes into contributions from the snow surface, snow volume, and ice surface, employing both least-squares fitting and Convolutional Neural Networks (CNN). Our findings challenge previous default assumptions, revealing that snow-volume scattering is as significant as that from the ice surface. To better characterize the snow scattering effects, we propose an adapted threshold first-maximum retracker algorithm (TFMRA) for CryoSat-2 radar freeboard retrieval, identifying a 70% threshold for ice floe retracking as the most accurate based on Operation IceBridge data. The algorithm, applied in CRYO2ICE campaigns and validated against snow buoy measurements, highlights the critical role of snow in altimeter retrievals. The result is particularly important given the recent unprecedented Antarctic sea ice loss and its application for future satellite missions like ESA's CRISTAL.

In this paper, a novel Reinforcement Learning (RL) algorithm designed for optimizing control in continuous action spaces is presented. The proposed method uses a "Target Interval Segmentation" (TIS) technique to discretize the action space into smaller segments, gradually narrowing these segments to facilitate a more efficient search for the optimal action. The method is evaluated on two important problems: the classical inverted pendulum as a benchmark and the problem of drug dosage control for leukemia patients. Simulation results indicate that the proposed method outperforms DQN, DDPG, TD3, SAC, and PPO algorithms in the drug dosage control problem. In the inverted pendulum benchmark, although SAC and TD3 delivered superior control performance, the proposed method exhibited significantly better stability. Unlike other algorithms, which experienced severe fluctuations in reward values, our method maintained consistent performance within a specific range, avoiding the large negative spikes that could jeopardize system stability. Overall, the simulation results suggest that this approach offers strong stability and accuracy, particularly in scenarios requiring precise control, and shows potential as an effective solution in continuous action spaces.

The human face is a silent communicator, expressing emotions and thoughts through it's facial expressions. With the advancements in computer vision in recent years, facial emotion recognition technology has made significant strides, enabling machines to decode the intricacies of facial cues. In this work, we propose ResEmoteNet, a novel deep learning architecture for facial emotion recognition designed with the combination of Convolutional, Squeeze-Excitation (SE) and Residual Networks. The inclusion of SE block selectively focuses on the important features of the human face, enhances the feature representation and suppresses the less relevant ones. This helps in reducing the loss and enhancing the overall model performance. We also integrate the SE block with three residual blocks that help in learning more complex representation of the data through deeper layers. We evaluated ResEmoteNet on three open-source databases: FER2013, RAF-DB, and AffectNet, achieving accuracies of 79.79%, 94.76%, and 72.39%, respectively. The proposed network outperforms state-of-the-art models across all three databases. The source code for ResEmoteNet is available at https://github.com/ArnabKumarRoy02/ResEmoteNet.

The inherent hysteresis nonlinearity of piezoactuated positioning stages (piezo-stages) is very difficult to deal with due to the amplitude-and frequency-dependent characteristics, which severely limits the tracking accuracy for high speed trajectories. In this work, we first develop a deep serial model to describe the dynamics of the piezo-stage by using historical voltage-displacement data over a period of time. It achieves relative prediction errors less than 0.16% on sinusoidal trajectories with frequencies greater than 72% of the resonance frequency of the piezo-stage through an elaborately designed network structure that includes direct connections between the input layer and the output layer. Then, we design an integral model predictive control (iMPC) and build a feedforward neural network (FNN) to learn its optimal solution offline. This forms the proposed FNN-iMPC and ensures the feasibility of evaluating the control law within the sampling time of 0.1ms. It achieves a maximum positioning error of 0.02µm for a ±32µm staircase reference signal and a maximum tracking error of 0.19µm for a sinusoidal reference signal with a range of 10µm and a frequency of 500Hz in real experiments.

This paper introduces Singular-value Gain Compensation (SGC), a robust preprocessing method for Ground Penetrating Radar (GPR) that integrates Singular Value Decomposition (SVD) and Time Gain Compensation (TGC). SGC effectively enhances the signal-to-noise ratio while maintaining weak signal integrity, facilitating the application of pretrained zero-shot segmentation models. Through extensive evaluations using simulated and real-world data, SGC demonstrates superior performance in image quality and segmentation accuracy compared to traditional methods. Optimizing the balance between noise reduction and signal enhancement significantly improves downstream tasks such as object segmentation. Additionally, SGC is computationally efficient, reducing both time and memory requirements, making it practical for large-scale infrastructure assessments. The method's efficacy in enhancing GPR image analysis without extensive computational resources marks a significant advancement in non-destructive subsurface inspection techniques.

To address the significant capital expenditures (CAPEX) and operating expenses (OPEX) associated with each antenna having its dedicated radio frequency (RF) chain in multiple-input multiple-output configuration, various techniques have been introduced, such as analog architectures, hybrid architectures, and index modulation (IM). Moreover, novel antenna architectures, such as graphene-based and dynamic metasurface antennas (DMAs), which leverage the unique properties of graphene and metamaterials, respectively, have been proposed to further reduce CAPEX and OPEX. In this study, we propose novel DMA-based IM schemes that aim to significantly reduce the power-hungry yet expensive RF chains, thereby improving the OPEX and CAPEX. Specifically, we propose a spatial modulation (SM) technique for the DMA architecture, which is achieved by activating either a single microstrip (microstrip-wise SM) with the help of a switching mechanism or a single DMA element (element-wise SM) by manipulating both the DMA reconfigurable weight and a switching mechanism. Our analysis shows that microstrip-wise SM achieves higher signal-to-noise ratios (SNRs) and, consequently, higher spectral efficiency compared with element-wise SM. However, in the high-SNR regime, the element-wise SM performs very close to the microstripwise SM while requiring lower complexity. Furthermore, we propose a generalized spatial modulation (GSM) technique for the DMA architecture, achieved by simultaneously activating multiple microstrips, to further enhance the spectral efficiency. We propose the design of DMA weights to maximize the SNR for the two considered SM schemes, obtaining closed-form solutions. For the GSM scheme, we develop an alternating algorithm to optimize the DMA weight and baseband precoder to maximize the spectral efficiency. Finally, we provide simulation results to validate our analysis and demonstrate the effectiveness of the proposed schemes.

The principle of maximum entropy is a well-established technique for choosing a distribution that matches available information while minimizing bias. It finds broad use across scientific disciplines and in machine learning. However, the principle as defined by Jaynes (1957) is susceptible to noise and error in observations. This forces real-world practitioners to use relaxed versions of the principle in an ad hoc way, negatively impacting interpretation. To address this situation, we present a new principle we call uncertain maximum entropy that generalizes the classic principle and provides interpretable solutions irrespective of the observational methods in use. We introduce a convex approximation and expectationmaximization based algorithm for finding solutions to our new principle. Finally, we contrast this new technique with two simpler generally applicable solutions theoretically and experimentally show our technique provides superior accuracy.

Automated translation systems are useful and essential tools for organisations operating in multilingual environments, offering significant advantages in efficiency and scalability. However, the challenge remains in optimising translation models for specific in-house tasks while maintaining high accuracy across diverse languages, particularly those with limited linguistic resources. Fine-tuning GPT-4 for multilingual translation presents a novel approach that addresses both high-resource and low-resource languages, providing a more adaptable solution for domain-specific translation needs. Through a comprehensive performance evaluation using automatic metrics such as BLEU, ROUGE, METEOR, and CHRF, the fine-tuned model demonstrated superior performance in high-resource languages, while revealing notable gaps in low-resource languages, indicating the need for further refinement. Additionally, the study emphasises the limitations of relying solely on automatic evaluation methods, as they may fail to capture subtle linguistic complexities. Despite these challenges, the findings suggest that fine-tuning offers substantial improvements in efficiency, accuracy, and scalability for in-house translation tasks, paving the way for more targeted applications in diverse organisational contexts.

Given the rapid emergence and applications of Large Language Models (LLMs) across various scientific fields, insights regarding their applicability in agriculture are still only partially explored. This paper conducts an in-depth review of LLMs in agriculture, focusing on understanding how LLMs can be developed and implemented to optimize agricultural processes, increase efficiency, and reduce costs. Recent studies have explored the capabilities of LLMs in agricultural information processing and decision-making. Nevertheless, a comprehensive understanding of the capabilities, challenges, limitations, and future directions of LLMs in agricultural information processing and application remains in its early stages. Such exploration is essential to provide the community with a broader perspective and clearer understanding of LLMs' applications, serving as a baseline for the current state and trends in the subject matter. To bridge this gap, this survey reviews the progress of LLMs and their utilization in agriculture, with an additional focus on 11 key research questions (RQs), where 4 RQs are general and 7 RQs are agriculture focused. By addressing these RQs, this review outlines the current opportunities and challenges, limitations, and future roadmap for LLMs in agriculture. The findings indicate that multi-modal LLMs not only simplify complex agricultural challenges but also significantly enhance decision-making and improve the efficiency of agricultural image processing. These advancements position LLMs as an essential tool for the future of farming. For continued research and understanding, an organized and regularly updated list of papers on LLMs is available at https://github.com/JiajiaLi04/ Multi-Modal-LLMs-in-Agriculture.

In this paper, we propose a thorough analysis of an advanced binary tree algorithm with multipacket reception (MPR) capabilities, along with timeliness metrics such as age of information (AoI) and packet loss probability (PLP) in a gated access setup scenario. We consider a finite number of users with Bernoulli arrivals and leverage packet preemption capabilities in their buffers, highlighting a novel and significant aspect. Initially, we explore the expected length of the collision resolution interval, conditioned on the anticipated number of users who have arrived beforehand. Notably, the preemption of user packets occurs both inside and outside of the corresponding collision resolution intervals. Subsequently, we illustrate the state transitions of contending users, conditioned on the MPR scenarios. Due to the capability of employing MPR, the receiver can successfully decode all the packets in an error-free channel. Furthermore, we demonstrate that, in gated access, the normalized throughput increases with an increase in the upper bound of MPR capabilities. Nonetheless, delay, PLP, and AoI exhibit superior performance in the unnormalized case. This demonstrates that enhancing physical resource capabilities in the unnormalized case leads to improved performance in a random-access channel when using binary tree algorithms.

Printed electronics using flexible substrates are an emerging area, allowing next generation electronics that can conform and flex with different surfaces, from human skin to in clothing. In the hybrid integration, or sea-of-rigids approach, conventional microchips are mounted onto (generally) plastic substrates such as polyethylene naphthalate (PEN), with (typically) printed silver tracks for interconnections between components. An ongoing research direction is to replace plastic substrates with biodegradable substrates, and to replace silver tracks with non-heavy metal-based tracks. While the substrates and tracks form only part of any overall system, replacing them is a step towards increased sustainability and helps to meet net-zero goals for printed electronic systems. Previously, several papers have investigated printed carbon tracks for low-frequency analog sensing applications. This paper explores the feasibility of using printed carbon tracks on biodegradable substrates for high-frequency sensing applications such as digital signaling over a Serial-Peripheral Interface (SPI). We investigate the printability, thermal stability, and electrical conductivity of carbon ink screen-printed onto six commercially available sustainable and flexible substrates. This includes modeling the connection formed and testing with SPI communications over a range of clock frequencies. We explore multi-layer screen printing to reduce the electrical resistance of the carbon tracks enabling their use in high-frequency SPI communication. This method was evaluated successfully in a custom experimental setup composed of two digital electronic systems connected through a carbon-based flexible connector. Connectors made through two layers of carbon prints allowed short distance SPI communication at a clock frequency of 1 MHz. Connectors of three or more carbon prints allowed SPI communication at clock frequencies up to 16~MHz. The NatureflexTM substrate showed the best overall trade-off in printing versus electronic performance among the sustainable substrates. The results of this study provide guidelines for materials selection and device fabrication for printing onto emerging biodegradable substrates to be used effectively in digital electronic systems.

The low Earth orbit (LEO) satellite megaconstellation can provide ubiquitous coverage and high-performance connectivity, supporting multi-slice applications with various key performance indicator (KPI) requirements. However, due to the dynamic nature of LEO satellites, limited resources, and the diverse demands of different slices, managing user association (UA) and resource allocation becomes an increasingly challenging task in areas with overlapping satellite coverage. This paper proposes a joint optimization model for UA and resource allocation in multi-slice LEO satellite networks (SLSNs). Based on mixed-integer non-linear programming (MILP), our model minimizes the total propagation delay and optimizes the demand satisfaction ratio (DSR) using a Max-Min approach to ensure each slice meets its unique throughput requirements. In addition, a visibility-aware component is incorporated to prioritize longer satellite visibility, reduce handovers, and improve network stability. Due to the computational complexity of the MILP model, we propose a heuristic-based balanced association with delay-aware bandwidth distribution (B-DAD) approach. B-DAD operates in two phases: the initial UA phase selects satellites based on a combined metric of delay, load, and visibility duration, while the residual bandwidth distribution phase reallocates unused bandwidth among associated users proportionally. Extensive simulations demonstrate that our approaches significantly improve DSR, propagation delays, transmission delays, and network stability compared to the widely adopted benchmark maximum sum of data rate (Max-SR) method under varying elevation angles. Our findings highlight the effectiveness of the MILP model in achieving optimal solutions and the efficiency of B-DAD as a scalable alternative for large-scale scenarios.