Generating a concise and informative video summary from a long video is important, yet subjective due to varying scene importance. Users' ability to specify scene importance through text queries enhances the relevance of such summaries. This paper introduces an approach for query-focused video summarization, aiming to align video summaries closely with user queries. To this end, we propose the Fully Convolutional Sequence Network with Attention (FCSNA-QFVS), a novel approach designed for this task. Leveraging temporal convolutional and attention mechanisms, our model effectively extracts and highlights relevant content based on user-specified queries. Experimental validation on a benchmark dataset for query-focused video summarization demonstrates the effectiveness of our approach.

For safety and robustness of AI systems, we introduce _topological parallax_ as a theoretical and computational tool that compares a trained model to a reference dataset to determine whether they have similar multiscale geometric structure. Our proofs and examples show that this geometric similarity between dataset and model is essential to trustworthy interpolation and perturbation, and we conjecture that this new concept will add value to the current debate regarding the unclear relationship between "overfitting"' and "generalization'' in applications of deep-learning. In typical deep-learning applications, an explicit geometric description of the model is impossible, but parallax can estimate topological features (components, cycles, voids, etc.) in the model by examining the effect on the Rips complex of geodesic distortions using the reference dataset. Thus, parallax indicates whether the model shares similar multiscale geometric features with the dataset. Parallax presents theoretically via topological data analysis [TDA] as a bi-filtered persistence module, and the key properties of this module are stable under perturbation of the reference dataset.

Unsustainable fishing practices worldwide pose a major threat to marine resources and ecosystems. Identifying vessels that do not show up in conventional monitoring systems---known as ``dark vessels''---is key to managing and securing the health of marine environments. With the rise of satellite-based synthetic aperture radar (SAR) imaging and modern machine learning (ML), it is now possible to automate detection of dark vessels day or night, under all-weather conditions. SAR images, however, require a domain-specific treatment and are not widely accessible to the ML community. Maritime objects (vessels and offshore infrastructure) are relatively small and sparse, challenging traditional computer vision approaches. We present the largest labeled dataset for training ML models to detect and characterize vessels and ocean structures in SAR imagery. xView3-SAR consists of nearly 1,000 analysis-ready SAR images from the Sentinel-1 mission that are, on average, 29,400-by-24,400 pixels each. The images are annotated using a combination of automated and manual analysis. Co-located bathymetry and wind state rasters accompany every SAR image. We also provide an overview of the xView3 Computer Vision Challenge, an international competition using xView3-SAR for ship detection and characterization at large scale. We release the data (\href{https://iuu.xview.us/}{https://iuu.xview.us/}) and code (\href{https://github.com/DIUx-xView}{https://github.com/DIUx-xView}) to support ongoing development and evaluation of ML approaches for this important application.

With the advent of the Internet, the gaming industry has grown tremendously in business, which also raises concerns for cheating and unfair gameplay. In this paper, we propose a novel approach (GiNA) using Blockchain technology to address a few problems with online Peer-to-Peer (P2P) games. GiNA uses two different data packet transfer schemes to ensure the security and authenticity of the data packet sent and received by game clients. More sensitive data uses a Smart contract-based ON-CHAIN data packet transfer solution and less sensitive data uses an OFF-CHAIN data packet transfer solution with end-to-end encryption for data security. A marketplace where peers can buy and sell purchasable assets with the help of Gicoins. Gicoins is a stable token with compliance with the ERC 20 token of Etheruem Blockchain. Later, a low cost and low bandwidth utilization data storage solution is proposed for storing data in a decentralized and distributed manner. Results show that the performance of the proposed approach GiNA is better in comparison to the traditional approaches with parameters such as latency, scalability, packet loss percentage, Blockchain (BC) performance, and data storage comparison.

The standard voltage boosting DC-DC converters may present a monotonically increasing or decreasing voltage gain. Nevertheless, to achieve significantly escalated voltage output, it is essential to employ voltage multiplier cells (VMC). Driven by this feature, a modified non-isolated ultra-lift luo converter (MULLC) based on VMC is adopted in this paper, referring to as MULLC-VMC. Unlike the standard ULLC, the proposed converter exhibits various distinguishing features like lower switch stress on the main power switch, zero-current switching, and suppression of electro-magnetic induction. Consequently, this allows a selection of power switch with low on-state resistance and low voltage rating that eventually helps in reducing the conduction losses. Proceeding further, an emphasis has been given to provide detailed description concerning to operating principle and different operating modes of the proposed converter. Besides, a steady-state analysis of the devised topology under continuous conduction mode alongside design procedure is expressed. Furthermore, a detailed comparative analysis is performed to demonstrate the superiority of the proposed converter over similar state-of-the-arts converters. Finally, stability and operation of the proposed MULLC-VMC is confirmed through frequency response analysis and MATLAB/Simulink studies, respectively.

Ridesharing is an effective, affordable and environment-friendly way of transportation. This paper proposes a Hybrid Genetic Algorithm(HGA) using elements from Simulated Annealing(SA) and Local search(LS) which is very suitable for Ridesharing related applications. The designed algorithm efficiently handles advanced constraints like timing windows for pick-up, detour time(or distance), and waiting-time minimization.

Personal computers (PCs) and many other household appliances employ multi-port switched-mode power supply (SMPS). However, several problems have been encountered with conventional SMPS which typically includes degraded power factor (PF), sluggish dynamic response, and high source current harmonics. This will potentially lead to power quality (PQ) issues. In order to confront these issues, this article proposes a modified magnetic energy recovery switch (MERS), which is employed at the front end of the conventional SMPS for power factor correction (PFC) while alleviating the PQ issues. In the present work, a detailed designing of magnetics including an inductor, capacitors, and transformers are thoroughly discussed. Simulation is conducted in MATLAB/Simulink environment for both conventional and proposed PFC SMPS. Further, a laboratory prototype is developed, and results are recorded, which are found to be in line with the simulation performance. Moreover, the performance of the proposed topology is investigated under different source voltage and loading condition. A comparison between conventional SMPS and proposed one intensely brings out the advantages of the proposed modified MERS based PFC approach in multi-port SMPS application.

Solar photovoltaic (SPV) power generating systems requires a robust and reliable control to perform the multitasking operation under different operating conditions. This paper presents an application of new square-error regularized NLMS (normalized least mean square) algorithm for integrating SPV system to the utility grid. Here, variable step-size regularization is adopted in order to fix the problem associated with fixed regularization including high steady state error and low convergence speed. This adaptive time varying parameter is designed by taking the inverse of weighted square error and implemented to the fundamental weight extraction mechanism for generating the inverter reference current. Subsequently, the proposed control algorithm enables this inverter to perform the multiple objectives including grid current harmonic alleviation, reactive current support, power factor correction and load balancing without affecting their major task of real power injection. Moreover, this control ably regulates the DC-bus voltage within the acceptable limits to maintain good dynamic response under transient conditions. The adopted two-stage grid connected SPV system with proposed control has been modelled using MATLAB/Simulink software for performing simulation study. Further these simulation results are verified through laboratory developed prototype using waijung environment of ARM cortex M4 microcontroller.