The Quantum Key Distribution (QKD) method leverages the principles of quantum mechanics to securely create and distribute private keys using quantum systems and an authenticated public classical channel. Despite offering informationtheoretical security, its physical implementations often suffer from unintended information leakage, known as side channels, which eavesdroppers can exploit to obtain the private key. This study introduces a classical-side-channel attack on the privacyamplification step of a general QKD protocol based on matrix hashing, utilizing machine-learning techniques to analyze powerconsumption leakage. Through multiple simulated scenarios, the study found that the gradient-boosting machine consistently outperformed other models, recovering the entire private key for high measuring-instrument sampling rates, regardless of the hashing matrix size and noise level tested. Additionally, the study proposes a strategy based on analyzing the confusion matrices of the model to facilitate a brute-force search for the critical feasible in the case of a non-perfect model. Furthermore, the study discusses countermeasures such as noise insertion, masking, and randomization techniques, potentially restoring the QKD protocol's information-theoretical security. This work demonstrates that machine-learning techniques can effectively characterize the leakages in a QKD implementation and create potent attacks.

This paper examines the development of community microgrids (CMGs) as a strategy to bridge the gap in energy equity in underserved communities that have long faced energy and environmental injustices. By introducing three novel equity-oriented indices-the Community Energy Financial Index (CEFI), the Community Energy Resiliency Index (CERI), and the Community Energy Sustainability Index (CESI), we aim to provide a framework to evaluate energy-related inequalities and challenges within communities. The proposed framework quantifies how the deployment of CMGs can systematically address obstacles that community members face in accessing clean and reliable energy affordably. We then propose a two-stage stochastic mixed-integer programming model, utilizing Benders decomposition, to optimize the planning and operation of CMGs in a community context in the face of operational uncertainties. Through a case study based on three energy-poverty neighborhoods in the Greater Houston area, our findings reveal, for the first time, that significant improvements in energy equity-related metrics-such as affordability, cost-effectiveness, resilience, and sustainability-can be achieved for community members through targeted investment and strategic planning of CMGs. Furthermore, our comprehensive analysis shows that under various budget scenarios and technology selections, CMGs planned with equity in mind consistently outperform the business-as-usual case and the conventional approach which focuses solely on cost minimization, underscoring the potential of CMGs to foster equitable, resilient, and sustainable energy systems for the future.

Proficiency in problem-solving techniques is essential for engineers across various disciplines. Achieving a successful answer to a specific problem in sciences often necessitates a knowledge and understanding of aspects of the problem, and a corresponding optimal solution based on available problem-solving strategies. However, a need exists for a foundational framework, i.e., a basic model that incorporates generic visual elements within the realm of problem-solving. This model should avoid delving into unnecessary specifics of particular applications while addressing broader issues effectively. First, this paper comprehensively reviews current problem-solving theories to spot merits and gaps. Second, drawing inspiration from the idea of diversity, inclusion and equality in the natural organisation of humans, we address the crucial issue by conceptualising an understandable framework termed Diversification-Agentsourcing-Conversification, or abbreviated DAC theory. Our abstract model is derived from finding shared patterns of solutions in specific successful solvers. The components constituting the DAC model are: diversifier; agents; conversifier; and, their interconnections-plusprotocol, abbreviated interconnopol. The diversifier first parses a query problem among a set of networked intelligent agents; the agents, as partial decision-makers, play complementary roles. Afterwards, their responses are directed towards a conversification. The unique conversifier, taking into account all aspects, ultimately generates a solution to the problem through a frequently intricate and nonlinear process of agents decisions synthesis. We provide underlying theoretical facts and manifest how diverse principles and methodologies can be simply elucidated by our universal model. It is especially effective for complicated decision-making situations in the presence of uncertainty, where success is crucial. The key advantages of the DAC model are twofold: firstly, it demystifies existing solutions; and secondly, it plays a crucial role in transferring the knowledge acquired from the general solution pattern to effectively address new questions.

A document by Fan Gao. Click on the document to view its contents.

This paper presents a new method to characterize the linearity of on-chip filters, considering the interaction between cascaded blocks. Unlike conventional methods, this approach uses only one buffer, simplifying the design and improving accuracy. This method accurately extracts the filter's transfer function in both its bandpass and stopband. The filter and buffer are designed using GlobalFoundries 22nm FDX technology, incorporating a back-gate biasing tuning mechanism in the buffer design. Results show a variation of ≈ 0.17 V and ≈ 0.19 V in the threshold voltage of PMOS and NMOS devices, respectively, for |1.2 V | shift in their back-gate voltage, aiming to maintain the performance of the buffer under process variation. The post-layout simulations demonstrate that the new method achieves a filter linearity of IIP 3 = 10.46 dBm, with an accuracy of 99.4% compared to the standalone filter's linearity. Similar consistency is observed across various process corners.

Nano and microsatellites have expanded the acquisition of satellite images with higher spatial, temporal, and spectral resolutions. Nevertheless, downlinking all this data to the ground for processing becomes challenging as the amount of remote sensing data rises. Custom onboard algorithms are designed to make real-time decisions and to prioritize and reduce the amount of data transmitted to the ground. However, these onboard algorithms frequently require cloud-free bottom-ofatmosphere surface reflectance estimations as inputs to operate. In this context, this paper presents DTACSNet, an onboard cloud detection and atmospheric correction processor based on lightweight Convolutional Neural Networks. DTACSNet provides cloud and cloud shadow masks and surface reflectance estimates 10x faster than the operational Sentinel-2 L2A processor in onboard hardware: 7 mins vs. 1 hour for a 10,980 x 10,980 scene. The DTACSNet cloud masking, based on a lightweight neural network, obtains the highest F2-score (0.81), followed by the state-of-the-art KappaMask (0.74), Fmask (0.72) and Sen2Cor v.2.8 (0.51) algorithms. Additionally, validation results on independent datasets show that DTACSNet can efficiently replicate Sen2Cor surface reflectance estimates, reporting a competitive accuracy with errors below 2%. Code is available at https://github.com/spaceml-org/DTACSNet.

In the burgeoning field of intelligent transportation systems, the integration of Generative Artificial Intelligence (AI) into vehicular networks presents a transformative potential for the automotive industry. This paper explores the innovative applications of generative AI in enhancing communication protocols, optimizing traffic management, and bolstering security frameworks within vehicular networks. By examining current technologies and recent advancements, we identify key challenges such as scalability, real-time data processing, and security vulnerabilities that come with AI integration. Additionally, we propose novel applications and methodologies that leverage generative AI to simulate complex network scenarios, generate adaptive communication schemes, and enhance predictive capabilities for traffic conditions. This study not only reviews the state of the art but also highlights significant opportunities where generative AI can lead to groundbreaking improvements in vehicular network efficiency and safety. Through this comprehensive exploration, our findings aim to guide future research directions and foster a deeper understanding of generative AI's role in the next generation of vehicular technologies.

In this manuscript, the problem of estimating human vital signs through a plurality of single-input multiple-output frequency modulated continuous wave radars is investigated. A novel algorithm for real-time estimation of heart rate and respiration rate, and for apnea detection is developed. It employs a selection combining technique and a filtering method inspired by Bayesian filtering principles. Our experimental results evidence that the proposed algorithm, tested on commercial radar devices, outperforms well known vital signs estimation methods available in the technical literature at the price of a limited increase in computational complexity.

In light of the enormous increase in data generated each day, machine learning methodologies must be continuously improved and adapted to deal with ever-larger amounts of data. As an example of machine learning, we discuss association rule mining in this paper. To extract meaningful rules from large databases, several approaches have been developed. However, several metrics still need to be improved. Using serverless functions, this paper presents an implementation for distributed association rule mining. The Apollo-ARM orchestration implementation is based on the Apollo multi-cloud orchestration framework developed by the University of Innsbruck. Three contributions are made by this paper. First, we review existing algorithms and applications for parallel and distributed mining association rules. Second, we design and implement the Apollo-ARM implementation. Third, we compare Apollo-ARM with an Apache-Spark-based implementation in terms of the number of rules extracted, the quality of the rules, and the speed of the algorithm on various datasets. Based on the results of the experiments, Apollo-ARM was able to extract significantly more rules, with higher accuracy as well as significantly faster. As a result of our study, we argue that distributed association rule mining using serverless functions is a promising approach that should be further developed in the future.

As RISC-V processors become more widely distributed, security issues arise in them. To this end, security verification techniques for processors should be incorporated into the design process to ensure the processor specifications, security requirements, and its actual implementation are consistent. Among the verification techniques, assertionbased verification has emerged as one of the most promising techniques. Although assertions are widely used for functional verification, there is limited effort in applying assertions for security verification. Thus, in this work, a novel security assertionbased technique is introduced for verifying the invulnerability of the processors against attacks. The method works in three main steps, first, collecting the security properties related to RISC-V, second, systematically identifying characteristics of instruction set architecture for each security property. Third, with the guide of input from the previous step, a security-based assertion miner is proposed to automatically extract a set of security assertions for the RISC-V. The experiments show that the proposed method can automatically generate security assertions in a very short amount of time and detect all the inserted Hardware Trojans in the processor, thereby accurately verifying the security of the processor.

Knowledge graphs have become essential tools for structuring and representing complex information across various domains, enabling advanced applications such as semantic search, recommendation systems, and automated reasoning. The novel concept introduced in this paper involves enhancing multi-modal knowledge graph completion through fine-grained tokenization of discrete modality information, significantly improving the richness and utility of knowledge graphs. We modified the Mistral Large model to integrate and process diverse data modalities, including text, images, and audio, via separate encoders and a fusion layer for integrated representation. The modified model demonstrated superior performance in multi-modal knowledge graph completion tasks, as evidenced by improved precision, recall, and F1-scores compared to baseline models. The rigorous training regimen, which combined supervised and unsupervised learning techniques, optimized the model's ability to generate accurate and contextually rich knowledge graphs. Comprehensive evaluations highlighted the robustness, scalability, and generalization capabilities of the enhanced model, underscoring its potential for broad applicability in domains requiring sophisticated knowledge representation and reasoning capabilities.

Number Portability Databases (NPDB) play a pivotal role in modern telecommunications by facilitating the seamless transfer of phone numbers between service providers. However, traditional NPDB systems face significant challenges, including centralization, high costs, and security vulnerabilities. This paper proposes a novel approach leveraging blockchain technology through the ENUMER system. ENUMER, designed initially to enhance ENUM operations, demonstrates a unique suitability for NPDB tasks due to its decentralized architecture, enhanced security, and cost efficiency. Our study involved testing ENUMER on standard hardware to ensure results stem from architectural advantages rather than high-performance equipment. The findings reveal sub-millisecond retrieval times and rapid record updates, showcasing ENUMER’s potential to outperform traditional NPDB systems. Although direct comparison data from existing NPDBs is scarce, our results suggest that ENUMER can serve as both an immediate solution and a foundation for future NPDB developments. For example, integrating public key infrastructure for caller ID verification could further enhance its functionality. This research highlights ENUMER’s capability to address current NPDB limitations and offers a promising direction for future improvements in telecom infrastructure.

β function in continued fractions forms reality of Fermion-Dirac and Bose-Einstein statistics everytime stochastic distribution event happens. And entropy is decided everytime uniquely.

5G cellular systems are slowly being deployed worldwide delivering the promised unprecedented levels of throughput and latency to hundreds of millions of users. At such scales security is crucial, and consequently, the 5G standard includes a new series of features to improve the security of its predecessors (i.e., 3G and 4G). In this work, we evaluate the security of currently deployed 5G commercial networks in Europe and North America. Specifically, by collecting 5G signaling traffic in the wild in several cities in Spain, Germany, France, Canada, and the USA, we i) fact-check which 5G security enhancements are implemented in current deployments, ii) provide a rich overview of the implementation status of each 5G security feature in a wide range of 5G commercial networks in Europe and North America and compare it with previous results in China, iii) analyze the implications of optional features not being deployed, and iv) discuss on the still remaining 4G-inherited vulnerabilities. Our findings indicate that the rollout of 5G security features in commercial networks is still a work in progress. On the one hand, many networks continue to rely on 4G for their core network operations, which hinders the deployment of new security features (e.g., SUCI) and, on the other hand, fully-fledged 5G deployments lack mandatory security measures such as GUTI reallocation after paging. Moreover, we find that some operators fail to provide proper temporary identifier randomization, in both 4G and 5G networks. Some of the obtained results are aligned with results previously reported from China [1] and keep the European and North American studied networks vulnerable to some 4G attacks, during their migration period from 4G to 5G. Conversely, networks deployed in North America exhibit stronger adherence to 5G security standards, with near-complete compliance observed, in contrast to deployments in China and Europe where compliance levels are comparatively lower.

Ship detection algorithms in Synthetic Aperture Radar (SAR) imagery are broadly categorized into: (i) those that interpret sea clutter according to a pre-defined Probability Density Function (PDF), detecting anomalies in positions within the tail of the PDF based on a specified Probability of False Alarm (PFA); and (ii) those which utilize a sufficiently large training dataset to learn the decision boundary between the target and clutter classes. Despite numerous publications in both categories, a proper quantitative comparison of their performance is lacking. This study is a step towards crossing this chasm by conducting a direct comparison between two real-world representatives: (i) SUMO's K-distribution Constant False Alarm Rate (K-CFAR/SUMO) detector, and (ii) the Deep Learning Model that topped the xView3 (1 st DLM/xView3) challenge organized by the Defense Innovation Unit and Global Fishing Watch. The performance of both algorithms is characterized by tracking the number of False Alarms (FAs) and Missed Detections (MDs) in three labeled Sentinel-1A repeat-pass SAR images acquired in the Gulf of Guinea. The results demonstrate that 1 st DLM/xView3 outperforms K-CFAR/SUMO, achieving the best FAs-MDs trade-off.This paper is accepted for oral presentation in IEEE IGARSS 2024, Athens, Greece.

TINY TAPEOUT is a multi project chip platform that makes it easier and cheaper to get application specific integrated circuit (ASIC) designs manufactured. Open source tools and process design kits (PDKs [1]) are used so no restrictive licenses or non disclosure agreements (NDAs) are required. As the tools run on remote cloud servers no software needs to be installed locally on the user’s machine. As long as the template structure is followed, however, Tiny Tapeout can support the use of proprietary tools.

In the field of remote sensing, semantic segmentation of Unmanned Aerial Vehicle (UAV) imagery is crucial for tasks such as land resource management, urban planning, precision agriculture, and economic assessment. Traditional methods use Convolutional Neural Networks (CNNs) for hierarchical feature extraction but are limited by their local receptive fields, restricting comprehensive contextual understanding. To overcome these limitations, we propose a combination of transformer and attention mechanisms to improve object classification, leveraging their superior information modeling capabilities to enhance scene understanding. In this paper, we present SwinFAN (Swin-based Focal Axial attention Network), a U-Net framework featuring a Swin transformer as encoder, equipped with a novel decoder that introduces two new components for enhanced semantic segmentation of urban remote sensing images. The first proposed component is a Guided Focal-Axial (GFA) attention module which combines local and global contextual information, enhancing the model's ability to discern intricate details and complex structures. The second component is an innovative Attentionbased Feature Refinement Head (AFRH) designed to improve the precision and clarity of segmentation outputs through selfattention and convolutional techniques. Comprehensive experiments demonstrate that the accuracy of our proposed architecture significantly outperforms state-of-the-art models. More specifically, our method achieves mIoU (mean Intersection over Union) improvements of 1.9% on UAVid, 3.6% on Potsdam, 1.9% on Vaihingen, and 0.8% on LoveDA.

Large language models (LLMs) have garnered unprecedented advancements across diverse fields, ranging from natural language processing to computer vision and beyond. The prowess of LLMs is underpinned by their substantial model size, extensive and diverse datasets, and the vast computational power harnessed during training, all of which contribute to the emergent abilities of LLMs (e.g., in-context learning) that are not present in small models. Within this context, the mixture of experts (MoE) has emerged as an effective method for substantially scaling up model capacity with minimal computation overhead, gaining significant attention from academia and industry. Despite its growing prevalence, there lacks a systematic and comprehensive review of the literature on MoE. This survey seeks to bridge that gap, serving as an essential resource for researchers delving into the intricacies of MoE. We first briefly introduce the structure of the MoE layer, followed by proposing a new taxonomy of MoE. Next, we overview the core designs for various MoE models including both algorithmic and systemic aspects, alongside collections of available open-source implementations, hyperparameter configurations and empirical evaluations. Furthermore, we delineate the multifaceted applications of MoE in practice, and outline some potential directions for future research. To facilitate ongoing updates and the sharing of cutting-edge advances in MoE research, we have established a resource repository accessible at https://github.com/withinmiaov/A-Survey-on-Mixture-of-Experts.