Hyperspectral videos contain a larger number of spectral bands, providing extensive spectral information and material identification capabilities. This advantage confers hyperspectral trackers to achieve superior performance in challenging tracking scenarios. However, the limited availability of hyperspectral training data and the inability of existing algorithms to fully exploit hyperspectral information restrict the tracking performance. To address this issue, a novel framework, Spectral Prompt-based Hyperspectral Object Tracking (SP-HST), is proposed. SP-HST leverages a RGB tracking network as the main branch for feature extraction and tracking, which accounts for more than 98% of the total parameters and remains frozen during the training procedure. Additionally, the Spectral Prompt Learning (SPL) branch, comprising multiple lightweight prompt blocks, is introduced to generate complementary spectral representations as the prompt. The prompts contain abundant spectral information from hyperspectral data, enhancing the discriminative ability of features within the main branch. Furthermore, the Complementary Weight Learning (CWL) is employed to calculate the importance of spectral information from different prompts, enabling the features for hyperspectral object tracking to contain more spectral information that is absent in the feature of the main branch. By utilizing the spectral information as prompt, the number of trainable parameters is less than 2% of that in the tracking network, and the convergence is reached in 12 training epoch. Extensive experiments demonstrate the superiority of SP-HST, achieving a new state-of-the-art tracking performance, 71.3% of the AUC score on the HOTC dataset and 96.7% of the DP@20P score on the IMEC25 dataset. The code will be released at https://github.com/lgao001/SP-HST.

With the rapid development of large-scale Low Earth Orbit (LEO) satellite internet, Very Low Earth Orbit (VLEO) Earth observation constellations are increasingly using InterSatellite Links (LISLs) for cross-layer access to Internet satellites. It is an effective means to enhance data return throughput. When both observation and communication satellite constellations use lasers for networking, cross-layer access between the VLEO and LEO satellite networks requires reallocating lasers. This reallocation disrupts the original topology and impacts network performance. To maximize the collaborative operational efficiency of the double-layer network, we fundamentally analyze the relationship between topology links, throughput, and delay using graph theory. Innovatively, we discover the superiority of two axioms in addressing the cross-layer topology optimization problem, including the “Minimum Hop Count” and “Minimum Overlap Path”. Based on these axioms, a many-objective cross-layer topology optimization model is established that considers the hop count and the average link utilization frequency of both VLEO and LEO satellite networks. To reduce the reliance on centralized algorithms on global transmission demand, a Local Distributed Interaction Mechanism (LDIM) is proposed for cross-layer LISL establishment. An onboard Novel Distributed Many-Objective Topology Optimization (NDMTO) algorithm is also introduced for VLEO satellites to manage access strategies. Finally, we use real data from Typhoon LEKIMA to create a multi-task scenario and conduct packet-level simulations based on the Starlink and Dove constellations. The results indicate that, compared to existing benchmarks, the NDMTO algorithm improves data throughput by 26.73% and reduces the average transmission delay of emergency task data by 20.7%.

This work proposes a quantum-aided deep reinforcement learning (DRL) framework designed to enhance the accuracy of direction-of-arrival (DoA) estimation and the efficiency of computational task offloading in integrated sensing and communication systems. Traditional DRL approaches face challenges in handling high-dimensional state spaces and ensuring convergence to optimal policies within complex operational environments. The proposed quantum-aided DRL framework that operates in a military surveillance system exploits quantum computing’s parallel processing capabilities to encode operational states and actions into quantum states, significantly reducing the dimensionality of the decision space. For the very first time in literature, we propose a quantum-enhanced actor-critic method, utilizing quantum circuits for policy representation and optimization. Through comprehensive simulations, we demonstrate that our framework improves DoA estimation accuracy by 91.66% and 82.61% over existing DRL algorithms with faster convergence rate, and effectively manages the trade-off between sensing and communication and optimizing task offloading decisions under stringent ultra-reliable low-latency communication requirements. Comparative analysis also reveals that our approach reduces the overall task offloading latency by 43.09% and 32.35% compared to the DRL-based deep deterministic policy gradient and proximal policy optimization algorithms, respectively.

A document by Praveer Dubey. Click on the document to view its contents.

Ransomware attacks continue to escalate in frequency and sophistication, posing severe threats to global cybersecurity infrastructures. Recognizing the necessity for advanced detection methods, this research introduces a novel approach that leverages dynamic clustering and out-of-bound detection techniques to proactively identify ransomware activities in file system traffic. By integrating real-time data analysis and adaptive thresholding, the system ensures high sensitivity and specificity in detecting potential threats. The method's ability to adjust to evolving attack patterns and varied file system behaviors significantly enhances its practicality for robust cyber defense strategies. Results from extensive evaluations demonstrate that this approach not only accurately identifies ransomware signatures but also reduces false positives, making it a valuable tool for ensuring digital security.

Additive non-Gaussian noise processes and nonlinear channel characteristics are found to present a performance bottleneck for next-generation communication systems, since most classical links are currently designed with Gaussian and linear system assumption. The joint effect of these impairments not only make channel characterization difficult, but also make optimal receiver design non-trivial and complex. In such scenarios, the joint application of reproducing kernel Hilbert space (RKHS) based machine learning methods, and information-theoretic learning (ITL) criteria, allows for near-optimal mitigation of channel nonlinearity and non-Gaussian noise for upcoming beyond 5G systems. Across different deployment scenarios, RKHS based learning allows for hyperparameter independence, improves generalization in the low-data regime and complements other existing signal processing solutions such as deep learning. Also, existing ITL criteria based algorithms are found to deliver improved performance for machine learning over non-Gaussian additive noise scenarios. Furthermore, existing analytical frameworks for these RKHS based approaches readily allow for the extension of classical performance analysis methods; existing works report rigorous analytical results, and the near-optimality of RKHS based approaches. This paper highlights the benefits and the potential of hyperparameter free RKHS and ITL based methods for simple and generic mitigation of unknown underlying hardware impairments and noise processes.

This survey investigates the integration of artificial intelligence (AI) with reconfigurable intelligent surfaces (RISs), emphasizing its pivotal role in advancing wireless communications. Initially, it provides an overview of RIS variants-passive, active, and hybrid-and the evolution to beyond diagonal RIS (BD-RIS), detailing their architectures, operational modes, benefits, and challenges. The focus then shifts to various AI algorithms applied within the RIS context, including supervised, unsupervised, deep learning (DL), reinforcement learning (RL), deep reinforcement learning (DRL), federated learning (FL), graph learning (GL), transfer learning (TL), meta-learning techniques. We explore how AI enhances RIS functionality in wireless networks, optimizing key aspects such as channel estimation (CE), rate maximization, expanding coverage, energy efficiency (EE), and enhancing security within RIS-based networks. Additionally, the survey conducts a thorough performance analysis of these AI-enhanced RIS systems under different networks. Comprehensive summary tables are provided to elucidate system models, channel state information, AI algorithms, and optimization strategies. The survey concludes by synthesizing the insights gained from AI in RIS applications, identifying current challenges, and suggesting directions for future research, thereby underscoring AI's substantial impact on the evolution of RISs and the 6G and beyond wireless networks.

Generative artificial intelligence (GenAI) is a powerful technology that can leverage advanced AI algorithms to automatically create, manipulate, and modify valuable while diverse data. GenAI has recently attracted phenomenal attentions and can serve as a key supplement for empowering the other techniques, e.g., the emerging human digital twin (HDT) that requires high-fidelity virtual modeling and strong information interactions but possibly with scarce, biased and noisy data. This survey focuses on the implementation of GenAI enabled HDT over mobile networks, i.e., mobile GenAI-driven HDT, that particularly addresses human-centric applications, demanding customized service, user-friendly experience, context-aware response and seamless integration, while suffering from uncertain user mobility, time-varying internal/external factors, unstable end-to-end network performance, etc. We begin by elaborating the core requirements of human-centric applications, and then envision the potential of utilizing mobile GenAI-driven HDT. Following this, we delve into the fundamentals of mobile GenAI-driven HDT by discussing both GenAI and HDT in detail, and exploring how GenAI-driven HDT can be implemented under the end-edge-cloud collaborative framework over mobile networks. After that, we introduce several typical human-centric applications that may be well supported by mobile GenAI-driven HDT, including healthcare, manufacturing and intelligent transportation. Moreover, we analyze and review in-depth the technical issues and promising solutions for the realization of mobile GenAI-driven HDT, focusing on the large model deployment, personalized model evolution and immersive interactions over mobile networks. Finally, we conclude this survey by highlighting some open issues and future research directions.

The exploration of AI-generated creative content has increasingly focused on the ability of machine learning models to produce text that resonates with human-like creativity, particularly in the domain of poetry. The novel introduction of a multi-agent framework represents a significant advancement in enhancing the poetic capabilities of language models, allowing for the generation of poems that are not only technically proficient but also exhibit deep thematic coherence and creative expression. Through the collaborative efforts of specialized agents focusing on rhyme, meter, theme, and metaphor, the framework refines initial drafts generated by the core model, producing outputs that are structurally sound and rich in lexical and artistic depth. Experimental results demonstrated the framework's effectiveness in improving rhyme quality, rhythmic consistency, and overall creative merit when compared to standard baseline models, showcasing the potential of such a system to push the boundaries of AI-driven literary generation. The significance of this work lies in its contribution to the intersection of artificial intelligence and the arts, providing a sophisticated tool for exploring and expanding the creative capabilities of machine-generated text.

This project focuses on exploring the characteristics of the 74LS161 integrated circuit (IC) in a more specific set tailored for visual simulation. The exploration extends beyond theoretical analysis, incorporating practical experiments and simulations that highlight the IC's real-world features. Specifically, the project delves into timing analysis and the impact of load variations on the 74LS161's performance, which are crucial for understanding its operation in complex circuits. The article primarily discusses the principles of the 74LS161 IC digital logic, relevant features, and then applies the 74LS161 to a hexadecimal counter followed by simulations. The simulations are divided into two types: the first using NI Multisim software to generate circuit diagrams and perform static simulations, and the second involving code in Vivado Verilog to generate dynamic waveforms for simulation. The visual results of these two simulation methods are compared and evaluated against the actual properties and minor applications of the 74LS161 IC.

In an era where robots are increasingly becoming part of our daily lives, we still struggle to define exactly what constitutes a robot. Do robots deserve to be defined by biological markers? How can we differentiate between a mechanical android versus one with synthetic skin? Are robots simply advanced machines that serve as tools for humanity? Will they eventually join the fabric of human society as our agents, advisors, assistants, maids, nurses, and partners? Will they be our servants, equals or superiors? Will we merge with them to become human-robot hybrids? When robots become superintelligent will they be regarded as sentient? How can we even define sentience? What is a robot? Systema Robotica aims to answer these questions and serve as humanity's guide to a better understanding and coexistence with robots in a future of non-human superintelligences. The treatise defines the true nature of robots, classifies three evolutionary realms of intelligences, structures a definitive robot taxonomy, explores the role of robots within society, and introduces a novel definition and test for sentience.

An adaptive, a variable code rate,can respond to channel changes promptly channel coding method is proposed,a discrete memoryless binary source sequence 𝑿 is transformed losslessly to make it capable of error checking. After the transformation, 𝑿 has variable and fixed code rates through weighted probability arithmetic coding. Compared with a fixed code rate, a variable code rate can respond to channel changes promptly and effectively improve the transmission performance, thereby being suitable for more complex communication environments. We prove that when the code length is long enough, the decoding error probability of the weighted probability arithmetic coding is 0 and the encoding rate can reach 1. When the symbols in the sequence 𝑿 are uniformly distributed, the transmission rate can reach the channel capacity. In an additive-white-Gaussian-noise channel binary-phase-shiftkeying signal simulation experiment, the error correction performances under various fixed code rates were compared. When the fixed code rate was 0.5 and the block error rate (BLER) was 10 −3 the weighted probability arithmetic coding improved by 0.37 dB compared to that obtained using Polar CA-SCL, and when BLER was 10 −5 , it improved by 0.48 dB. The algorithm can be widely used in communication systems, data storage, and other fields.

Crowd simulation plays a crucial role in the prevention and management of urban public safety events. However, the complexity and heterogeneity of human behaviors pose significant challenges to achieving realistic and accurate simulations. Traditional rule-based physical models are limited by their reliance on singular physical formulas and parameters, making it difficult to effectively tackle the intricate tasks associated with crowd simulation. Recent research has introduced deep learning methods to tackle these issues, but most current approaches focus primarily on generating pedestrian trajectories, often lacking interpretability and failing to provide real-time dynamic simulations. To address the aforementioned issues, we propose a novel data-driven crowd simulation framework that integrates Physicsinformed Machine Learning (PIML) with navigation potential fields. Our approach leverages the strengths of both physical models and PIML. Specifically, we design an innovative Physicsinformed Spatio-temporal Graph Convolutional Network (PI-STGCN) as a data-driven module to predict pedestrian movement trends based on crowd spatio-temporal data. Additionally, we construct a physical model of navigation potential fields based on flow field theory to guide pedestrian movements, thereby reinforcing physical constraints during the simulation. In our framework, navigation potential fields are dynamically computed and updated based on the movement trends predicted by the PI-STGCN, while the updated crowd dynamics, guided by these fields, subsequently feed back into the PI-STGCN. Comparative experiments on two publicly available large-scale real-world datasets across five scenes demonstrate that our proposed framework outperforms existing rule-based methods in accuracy and fidelity. The similarity between simulated and actual pedestrian trajectories increases by 10.8%, while the average error is reduced by 4%. Moreover, our framework exhibits greater adaptability and better interpretability compared to methods that rely solely on deep learning for trajectory generation.

This paper introduces Humanized Agent-Based Models (h-ABM), a conceptual approach integrating Large Language Models (LLMs) into Agent-Based Models (ABMs) to enhance the realism of human behavior in simulations. Recent works have demonstrated the capabilities of LLMs for creating powerful agents; nevertheless, they don’t propose a framework for defining agents and guiding their integration into ABMs. Surveying previous work in the field, and from the findings of LLM agents, ABM literature, and cognitive frameworks, h-ABM proposes a modular framework for LLM agents in the context of ABMs. We finally compare the framework with other LLM agents’ proposals to see how they fit into i

Machine learning can be used as a systematic method to non-algorithmically "program" quantum computers. Quantum machine learning enables us to perform computations without breaking down an algorithm into its gate "building blocks", eliminating that difficult step and potentially reducing unnecessary complexity. In addition, the machine learning approach is robust to both noise and to decoherence, which is ideal for running on inherently noisy NISQ devices which are limited in the number of qubits available for error correction. Here we apply our prior work in quantum machine learning technique, to create a QGAN, a quantum analog to the classical Stylenet GANs developed by Kerras for image generation and classification. A quantum system is used as a generator and a separate quantum system is used as a discriminator. The generator Hamiltonian quantum parameters are augmented by quantum "style" parameters which play the role of the style parameters used by Kerras. Both generator parameters are trained in a GAN MinMax problem along with quantum parameters of the discriminator Hamiltonian. We choose a problem to demonstrate the QGAN that has purely quantum information. The task is to generate and discriminate quantum product states. "Real" product states are generated to be detected and separated from "fake" quantum product states generated by the quantum generator. The problem of detecting quantum product states is chosen to demonstrate the QGAN because is well known as purely quantum mechanical, has no classical analog and is an open problem for quantum product states of more than 2 qubits. With proper encoding of image pixels into quantum states as density matrices, the method demonstrated here for this quantum problem is applicable to GAN image generation and detection, the problem so successfully demonstrated by Kerras's classical GAN, but can be hosted on and taking advantage of the nature of quantum computers.

Aspect-based summarization has become increasingly important in scenarios where specific information needs to be extracted and condensed from vast amounts of text data, such as in e-commerce, healthcare, and finance. Introducing a novel approach that fine-tunes the Mistral Large model to target aspect-specific content, this research offers significant advancements in generating high-quality summaries that are both contextually accurate and domain-specific. The methodology involved a rigorous process of data collection, preprocessing, and aspect annotation, which collectively enhanced the model's ability to handle diverse linguistic structures and deliver precise summarization results. Extensive experiments conducted across multiple domains demonstrated the model's superior performance in key metrics such as precision, recall, and F1-score, highlighting its adaptability and effectiveness. The findings not only validate the robustness of the fine-tuned model but also reveal the potential for further optimization, particularly in complex domains. Through this research, a foundation has been laid for future innovations in generating aspect-based summaries, offering valuable insights for developing more targeted and sophisticated applications.

Bio-inspired whisker sensors are engineered for a range of applications, including fluid-flow sensing, environmental exploration, and texture analysis. While numerous existing sensor designs exist, current reviews primarily focus on use cases and transduction methods. However, a noticeable gap remains in developing durable whisker sensors that maintain consistent hysteresis, are easy to fabricate, and can be configured for different purposes. Our study compares four in-house designed whisker-inspired sensors to identify an efficient fabrication method, considering sensing schemes, compliance, and whisker elements. Comparative analysis of each configuration provides new insights into optimal whisker-based sensor design. Additionally, we evaluate the performance of these sensors through standard roughness measurements. The central goal is to fabricate a whisker sensor that is long-lasting and easy to build and reconfigure for various applications. Based on our observations, we recommend suitable elements for whisker sensor fabrication.

This paper presents a comprehensive review of the current research on QCD axions, focusing on their potential as a solution to the strong CP problem and as a candidate for cold dark matter. Through detailed Monte Carlo simulations, the distribution of axion masses and couplings is explored, revealing critical insights into the likelihood of various parameter values. The study also examines the role of higher-order corrections in the axion potential, which significantly influence the dynamics of axions in the early universe. Furthermore, the paper discusses the potential of ultralight axions as quintessence, offering an intriguing link between axion physics and the accelerated expansion of the universe. The detection prospects for axions through their coupling with photons are analyzed, with a focus on experimental approaches such as ADMX and CAST. Finally, the results from a recent study on axion physics within the framework of string theory are incorporated, providing a broader context for understanding the cosmological implications of axions. The findings underscore the importance of refining axion models and experimental techniques to enhance the understanding of these elusive particles.