Most NFTs (Non-Fungible Tokens) use multi-hop URLs to address the off-chain assets due costly on-chain storage, but path from underlying is fraught with instability, which may degrade its value. Hence, this paper aims answer question: Is NFT-to-Asset connection fragile? This makes a first step towards end by characterizing connections of 12,353 Ethereum NFT Contracts (6,234,141 in total) three perspectives, accessibility, and duplication. In order overcome challenges affecting measurement...

Particle swarm optimization (PSO) is a population‐based stochastic technique in smooth search space. However, category of trajectory problem with arbitrary final time and multiple control variables, the smoothness variables cannot be satisfied since linear interpolation widely used. In paper, novel Legendre cooperative PSO (LCPSO) proposed by introducing orthogonal polynomials instead interpolation. An additional variable introduced to transcribe original optimal fixed one. Then, practical...

Image watermarking is a technique for hiding information into images that can withstand distortions while requiring the encoded image to be perceptually identical original image. Recent work based on deep neural networks (DNN) has achieved impressive progression in digital watermarking. Higher robustness under various eternal pursuit of approaches. In this paper, we propose DB MARK, novel end-to-end framework boost DNN-based The key novelty synergy invertible (INN) and effective watermark...

Robust watermarking tries to conceal information within a cover imag e/video imperceptibly that is resistant various distortions. Recently, deep learning-based approaches for image have made significant advancements in robustness and invisibility. However, few studies focused on video using neural networks due the high complexity computational costs. Our paper aims answer this research question: Can well-designed be efficiently adapted watermarking? positive. First, we revisit workflow of...

Traffic lights control could be regarded as a multi-agent coordinated problem. A model-free reinforcement learning (RL) approach is powerful framework for solving such policy-making problems without prior environmental knowledge. In order to global policy, communication among agents needs built. To enable dynamic and scalable communication, we propose new RL model, CommNet based on Local Attention Mechanism (Attn-CommNet), which uses local selection attention mechanism between hidden layers...

This article solved the reentry optimization problem of RLV using Deep Reinforcement Learning-Deep Deterministic Policy Gradient (DDPG) method for continuous system decision making. Compared with traditional intelligent algorithm, DDPG algorithm trains appropriate action values each state value during flight by constructing neural network and critic network, avoiding problems caused improper segmentation algorithms. And through greedy process is prevented from falling into local optimum. By...

This paper investigates the problem of interplanetary transfer orbit optimization based on two intelligent algorithms, particle swarm (PSO) and biogeography-based optimization(BBO). In order to from Earth Mars orbit, a continuous thrust model is established normalized. The shooting method penalty function are used convert small space into nonlinear programming problem. By using algorithms optimize Earth-Mars trajectory problem, feasibility in verified, performance compared.

Pedestrian trajectory forecasting is a fundamental task in multiple utility areas, such as self-driving, autonomous robots, and surveillance systems. The future multi-modal, influenced by physical interaction with scene contexts intricate social interactions among pedestrians. mainly existing literature learns representations of deep learning networks, while the explicit patterns are not utilized. Different patterns, following or collision avoiding, will generate different trends next...

Evolutionary analog circuit design is a challenging task due to the large search space incurred by topology and device values. Applying genetic operators on randomly selected genes may make it difficult identify which part of sub-circuit beneficial evolution even destroy useful sub-circuits, potentially incurring stagnation evolutionary process bloat evolved circuits. In this paper, we propose tree-based approach called Shapley Circuit Tree that incorporates values for quantifying...

Federated learning (FL), integrating group fairness mechanisms, allows multiple clients to collaboratively train a global model that makes unbiased decisions for different populations grouped by sensitive attributes (e.g., gender and race). Due its distributed nature, previous studies have demonstrated FL systems are vulnerable poisoning attacks. However, these primarily focus on perturbing accuracy, leaving critical question unexplored: Can an attacker bypass the mechanisms in manipulate be...

Abstract To solve the problems of complicated long-distance cable laying process, low efficiency and easy damage, an intelligent distributed system based on a multi-machine coordination strategy is proposed in this paper. Firstly, cooperative control clamping push proposed, which solves problem complex laying. Secondly, cable-laying machine developed to realize perception adaptive field conditions. Finally, developed. The combination further strengthens cooperation stability whole provides...

The online data markets have emerged as a valuable source of diverse datasets for training machine learning (ML) models. However, from different providers may exhibit varying levels bias with respect to certain sensitive attributes in the population (such race, sex, age, and marital status). Recent dataset acquisition research has focused on maximizing accuracy improvements downstream model training, ignoring negative impact biases acquired datasets, which can lead an unfair model. Can...

It is very important of high-speed railway construction surveying and mapping. Especially the establishment survey control network searching aiming at layout, measurement, data processing adjustment calculation horizontal elevation network. The most part CP II CPIII mainly adopts second order leveling. Horizontal accuracy must to meet requirements high speed survey.

Web-based AI image generation has become an innovative art form that can generate novel artworks with the rapid development of diffusion model. However, this new technique brings potential copyright infringement risks as it may incorporate existing without owners' consent. Copyright quantification is primary and challenging step towards AI-generated traceability. Previous work only focused on data attribution from training perspective, which unsuitable for tracing quantifying in practice...

Time Delay Reservoir (TDR) can exhibit effects of high dimensionality and short-term memory based on delay differential equations (DDEs), as well having hardware-friendly characteristics. However, the predictive performance capacity standard TDRs are still limited, dependent hyperparameter oscillation function. In this paper, we first analyze these limitations their corresponding reasons. We find that reasons for such fused by two aspects, which trade-off between strength self-feedback...

Aiming at the flight control system with typical nonlinear time-varying characteristics, a design method of Linear Parameter Varying (LPV) based on intelligent optimization is proposed. Through method, constant PID robust controller sought to slowly varying model largest possible operating range, which makes both global stability. Firstly, kinematics modeling aircraft, LPV description carried out. Then designed for three channels and parameters are optimized by fireworks algorithm. Finally,...

NFTs (Non-Fungible Tokens) have experienced an explosive growth and their record-breaking prices been witnessed. Typically, the assets that represent are stored off-chain with a pointer, e.g., multi-hop URLs, due to costly on-chain storage. Hence, this paper aims answer question: Is NFT-to-Asset connection fragile? This makes first step towards end by characterizing connections of 12,353 Ethereum NFT Contracts (6,234,141 in total) from three perspectives, storage, accessibility duplication....

Robust watermarking tries to conceal information within a cover image/video imperceptibly that is resistant various distortions. Recently, deep learning-based approaches for image have made significant advancements in robustness and invisibility. However, few studies focused on video using neural networks due the high complexity computational costs. Our paper aims answer this research question: Can well-designed be efficiently adapted watermarking? positive. First, we revisit workflow of...

Recent advances in federated learning (FL) enable collaborative training of machine (ML) models from large-scale and widely dispersed clients while protecting their privacy. However, when different clients' datasets are heterogeneous, traditional FL mechanisms produce a global model that does not adequately represent the poorer with limited data resources, resulting lower accuracy higher bias on local data. According to Matthew effect, which describes how advantaged gain more advantage...

Image watermarking is a technique for hiding information into images that can withstand distortions while requiring the encoded image to be perceptually identical original image. Recent work based on deep neural networks (DNN) has achieved impressive progression in digital watermarking. Higher robustness under various eternal pursuit of approaches. In this paper, we propose DBMARK, novel end-to-end framework boost DNN-based The key novelty synergy invertible (INN) and effective watermark...