A document by Bisma. Click on the document to view its contents.

This article aims to address a crucial aspect raised by Berkan Dulek in his comments [1] on the original paper, "An Efficient Algorithm for Maneuvering Target Tracking"[2]

A document by wenhan li. Click on the document to view its contents.

MapReduce efficiently processes large data sets in a Petabyte/Exabyte scale distributed system. Applications such as reverse indexing, image and pattern recognition, and analytics rely extensively on this programming model for large-scale distributed data processing. MapReduce works by splitting the processing into two distinct phases: a Map phase, which distributes data processing, and a Reduce phase, which aggregates results. While open-source applications such as Apache Hadoop and its query engine "Hive" implement MapReduce, some challenges persist. Most crucially, high network I/O is incurred when Mappers write many rows. One solution (attempted by Apache Hive) is to pre-aggregate data at the Mapper and send this aggregated data to the Reducer. This attempts to reduce network I/O. However, this is insufficient for two reasons: first, Reducers are starved while Mappers aggregate data at Map-Side, and second, Mappers can run into memory overflow and need a re-compute if pre-aggregated results exceed their buffer capacity. We propose a design to solve these problems using an adaptive split size for Map-side pre-aggregation. By dynamically adjusting the size of the input data split for a Mapper, we ensure that (1) Reducers get a continuous stream of data without starving and (2) Mapper memory overflows are avoided, which also avoids the need to "flush and re-compute" earlier results.

We experimentally demonstrate waveguide and microring resonator (MRR) polarizers by integrating 2D graphene oxide (GO) films onto silicon (Si) photonic devices. The 2D GO films with highly anisotropic light absorption are on-chip integrated with precise control over their thicknesses and sizes. Detailed measurements are performed for the fabricated devices with different GO film thicknesses, coating lengths, and Si waveguide widths. The results show that a maximum polarization-dependent loss (PDL) of ~17 dB is achieved for the hybrid waveguides, and the hybrid MRRs achieved a maximum polarization extinction ratio (PER) of ~10 dB. We also characterize the wavelength-and power-dependent response for these polarizers. The former demonstrates a broad operation bandwidth of over ~100 nm, and the latter verifies performance improvement enabled by photo-thermal changes in GO films. By fitting the experimental results with theoretical simulations, we find that the anisotropy in the loss of GO films dominates the polarization selectivity of these devices. These results highlight the strong potential of 2D GO films for realizing high-performance polarization selective devices in Si photonic platform.

We demonstrate that data-driven system identification techniques can provide a basis for effective, non-intrusive model order reduction (MOR) for common circuits that are key building blocks in microelectronics. Our approach is motivated by the practical operation of these circuits and utilizes a canonical Hammerstein architecture. To demonstrate the approach we develop a parsimonious Hammerstein model for a nonlinear CMOS differential amplifier. We train this model on a combination of direct current (DC) and transient Spice (Xyce) circuit simulation data using a novel sequential strategy to identify the static nonlinear and linear dynamical parts of the model. Simulation results show that the Hammerstein model is an effective surrogate for the differential amplifier circuit that accurately and efficiently reproduces its behavior over a wide range of operating points and input frequencies.

This article explores how syslog contributes to IEC 62351 compliance, particularly in incident response scenarios. Additionally, it outlines the extended benefits of incorporating syslog into OT environments, beyond the confines of the IEC 62351 standard.

Secure software development is essential for safeguarding systems against an increasing array of cyber threats. Evaluating the capabilities of advanced language models in providing secure coding advice is crucial for understanding their potential as reliable tools for developers. The study benchmarks ChatGPT and Google Gemini across metrics such as accuracy, relevance, comprehensiveness, and adherence to best practices, offering a detailed comparative analysis of their performance. Google Gemini demonstrated superior accuracy and adherence to secure coding standards, while ChatGPT excelled in generating contextually relevant and comprehensive responses. The findings highlight the strengths and weaknesses of each model, providing insights into their application in secure software development. The study's results inform developers on leveraging these models to enhance their coding practices, contributing to the creation of more secure and resilient software systems.

This work considers the design of a generalized Gray-mapping process to multilayer multicast non-orthogonal multiple access (NOMA) transmission with arbitrary modulation orders. Unlike orthogonal multiple access, joint-multilayer Graymapping can provide significant energy savings and bit error rate (BER) improvements, which can be used to alleviate the degradation caused by the inherent multilayer interference in NOMA. The obtained improvement is due to the increased Euclidean distance that Gray-mapping provides for certain layers/symbols. To evaluate the impact of Gray-mapping, closed-form expressions are derived for the exact BER with imperfect successive interference cancellation. The obtained analytical and simulation results demonstrate that the proposed scheme can offer up to 10 dB gain and 94% energy saving compared to conventional NOMA in certain scenarios, where such a performance gain can be shared between the layers by selecting an appropriate power assignment. Moreover, feasibility maps are generated to demonstrate the additional flexibility that Gray-mapping can offer in terms of quality of service satisfaction for the various layers at lower signal to noise ratios.

Asymptotic signals have been introduced to approximate integrals with the stationary phase approximation. However, the deviation between integrals and their approximations is not bounded without further constraints. We propose an alternative with the introduction of a parameter defined in the complex phase representation. The magnitude of this new parameter controls the local harmonicity of the signal. It therefore controls the deviation between integrals and their approximations, which do not rely on the stationary phase principle. Approximations of Fourier transforms illustrate the results. The parameter is also involved in additional properties of diverse nature.

The pansharpening task has always been a focal point for researchers. However, existing methods have yet to fully exploit the modality correlation between Panchromatic (𝑃 𝐴𝑁) and Multispectral (𝑀𝑆) images, leading to the loss of spatial high-frequency information and spectral distortion during fusion. Additionally, they apply the same processing standards to information at different scales, causing these pieces of information to mix and interfere. To address these issues, we propose a multi-scale fusion network based on a continuous information guidance mechanism named CIGformer. We focus on improving the interaction mechanism between 𝑃 𝐴𝑁 and 𝑀𝑆. First, we introduce an Intensity Substitute Block (ISB) to decompose the shared and unique information of 𝑃 𝐴𝑁 and 𝑀𝑆, initializing the subsequent information guidance. Then, we construct an Information Guidance Block (IGB) based on the Transformer architecture, which is the core of our information guidance. It enables adaptive balanced retention of the unique information of 𝑃 𝐴𝑁 and 𝑀𝑆 without affecting the utilization of their shared information. Furthermore, we build a multi-level Encoder-Decoder bidirectional pyramid structure to mitigate the impact of multi-scale information intermixing. At each encoder level, we continuously use IGB to achieve optimal information utilization. Finally, we introduce a consistency loss function to evaluate the retention of unique information, assisting in the training process. Experiments on the GaoFen-2 and WorldView-3 datasets show that based on unique information guidance, our proposed interaction method significantly enhances the utilization efficiency of 𝑃𝐴𝑁 and 𝑀𝑆. Our code is available at https://github.com/Xidian-AIGroup190726/CIGformer.

Offline reinforcement learning (RL) learns policies for solving sequential decision problems directly from offline datasets. Most existing works focus on countering out-of-distribution (OOD) behavior to improve decision-making. This work investigates the fundamental Bellman inconsistency problem that leads to suboptimal decision-making, which is an essential factor causing suboptimality in offline settings. An offline RL algorithm without regularization and pessimism (RFORL) is proposed. The RFORL algorithm constrains uncertainties constructed by the inconsistency among Bellman estimations under an ensemble of learned Qfunctions. Compared to existing offline RL methods, the proposed RFORL achieves exponential convergence and allows learning the best policy from the given offline dataset, without adopting regularization techniques or pessimism. We empirically demonstrate that the Bellman inconsistency constraint mechanism is effective in improving policy learning. Consequently, RFORL outperforms existing offline RL approaches on most tasks of the D4RL benchmark.

The robust unsupervised framework of clustering models is essential in numerous machine learning tasks due to its ability to identify hidden relations between samples, resulting in a comprehensive understanding and interpretation. Many datasets contain samples that naturally cannot link to each other, but the degree of similarity between them makes it almost impossible for clustering models to distinguish the differences; this is called the overlapping issue. However, integrating support vector machines, feature selection, and dimensional reduction techniques with clustering models might still be incapable of providing an optimal solution. As a result, it adversely affects performance and leads to inconsistent partitioning, unreasonable interpretation, and false confidence. This study addresses these issues by proposing a novel unsupervised data separation equation that is based on the concepts of tension and separation gained by finding the cannot-link relations based on cluster centroids. The equation validated in diverse scenarios to demonstrate its ability to improve outcomes. The experimental results prove that the proposed equation assists in reducing the reliance on parameter tuning and constraints, thereby enhancing performance and effectively addressing the challenge of outliers.

The following material is devoted to the generalization of the chaos modeling to random fields in communication channels and its application on the space-time filtering for the incoherent paradigm. The approach is based on the "Markovian" trend in modeling of random fields,and it is applied to chaos field modeling through the well-known concept of the random "treatment" of deterministic dynamic systems, first presented by A. Kolmogorov, M. Born, etc. The material applies the generalized Stratonovich-Kushner Equations (SKE) for the optimum filtering of chaotic models of random fields and its simplified quasioptimum solutions. In addition to this, the application of the multi-moment algorithms for quasi-optimum solutions is considered and, it is shown that for scenarios, when the covariation interval of the input random field is less than the distance between the antenna elements, the gain of the space-time algorithms against their "time" analogies is really significant.

In this paper, we raise five questions and find solutions: What are the positive and negative effects of broadband Internet? Which type of Internet access is preferable? Which one of them needs more attention? How much and what kind of developments affect the Internet? What solutions are there to improve the current situation? First, fourteen advantages and positive effects of broadband Internet in every corner of life are presented. Second, fifty criteria are extracted for comparing wired and wireless broadband Internet. The third task, they are compared by describing the characteristics of optical fiber-based fixed and fifth-generation-based fixed/mobile broadband Internet access. Fourth, issues that directly and/or indirectly affect the global development of broadband Internet are outlined. The fifth task elicits twenty items on balanced sustainable development and strategic planning of broadband Internet. Finally, we propose ten solutions to improve the current state of global broadband Internet, reduce the digital divide in the world, and decrease the negative social effects of broadband Internet.

The emergence of Industry 4.0 has significantly transformed manufacturing landscapes, introducing interconnected, dynamic, and data-rich environments. This paper focuses on the application of Industrial Machine Learning (I-ML) within these evolving manufacturing contexts, exploring both the challenges and future prospects of its integration. A systematic literature review, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, forms the foundation of our analysis, characterizing the role of ML in modern manufacturing, its current challenges, and future trends. This research delves into the implications of I-ML in various manufacturing scenarios, such as predictive maintenance, anomaly detection, and quality control, providing a comprehensive overview of its practical applications. We also address the critical need for sustainable, reproducible, and reliable performance in industrial applications and explore strategies for overcoming barriers to ML adoption in the industry. The paper discusses recommendations for future research directions, aiming to bridge the gap between ML advancements and their practical, scalable implementation in industrial settings. Through this work, we aim to contribute to the identification of challenges and future research directions to the ongoing digital transformation of manufacturing industries, offering insights into how ML can be effectively leveraged in the era of Industry 4.0 (I4.0).

Background: Document forgery poses a significant challenge to the integrity of sensitive documents and necessitates the development of advanced anti-forgery techniques. This study explored the integration of piezoelectric nanomaterials, specifically zinc oxide (ZnO) nanowires and barium titanate (BaTiO3) nanoparticles, into document fibers for enhanced forgery detection. Methods: This review examines the superior mechanical properties of ZnO nanowires and the high dielectric constant, piezoelectricity, and ferroelectricity of BaTiO3 nanoparticles, making them suitable for embedding within document fibers. Advanced scanning probe techniques, such as piezoresponse force microscopy (PFM), scanning Kelvin probe microscopy (SKPM), and electrostatic force microscopy (EFM), have been used to characterize the electrical properties and structural integrity of nanomaterials. Results: Embedding piezoelectric nanomaterials within document fibers enables the detection of tampering through changes in electrical properties, providing a non-invasive and highly sensitive forgery detection method. This study highlights the importance of optimizing the manufacturing processes, developing rigorous measurement protocols, and exploring alternative piezoelectric nanomaterials. Conclusions: The integration of piezoelectric nanomaterials into document fibers is a promising approach for addressing the persistent challenges of document forgery. Future research should focus on refining manufacturing techniques, improving measurement accuracy, and expanding the applications to various sensitive documents.

Large language models (LLMs) with billions of parameters and trained on massive amounts of crowdsourced public data have made a dramatic impact on natural language processing (NLP) tasks. Domain specific 'finetuning' of LLMs has further improved model behavior through task-specific alignment and refinement. However with the widespread development and deployment of LLMs, existing vulnerabilities in these models have made way for perpetrators to manipulate them for malicious intentions. The widespread integration of LLMs in clinical NLP underscores the looming threat of privacy leakage and targeted attacks like prompt injection or data poisoning. In this work, we designed a systematic framework to expose vulnerabilities of clinical generative language model with a specific emphasis on its application to clinical notes. We design three attack pipelines to highlight model's susceptibility to core types of targeted attacks-(i) instruction-based data poisoning, (ii) trigger-based model editing and (iii) membership inference on de-identified breast cancer clinical notes. Our proposed framework is the first work to investigate the extent of LLM-based attacks in the clinical domain. Our findings reveal successful manipulation of LLM behavior, prompting concerns about the stealthiness and effectiveness of such attacks. Through this work, we hope to emphasize on the urgency of comprehending these vulnerabilities in LLMs and encourage the mindful and responsible usage of LLMs in the clinical domain.