Partial multi-label learning（PML） allows learning from rich-semantic objects with inaccurate annotations, where a set of candidate labels are assigned to each training example but only some of them are valid. Existi...
详细信息
Partial multi-label learning（PML） allows learning from rich-semantic objects with inaccurate annotations, where a set of candidate labels are assigned to each training example but only some of them are valid. Existing approaches rely on disambiguation to tackle the PML problem, which aims to correct noisy candidate labels by recovering the ground-truth labeling information ahead of prediction model induction. However, this dominant strategy might be suboptimal as it usually needs extra assumptions that cannot be fully satisfied in real-world scenarios. Instead of label correction, we investigate another strategy to tackle the PML problem, where the potential ambiguity in PML data is eliminated by correcting instance features in a label-specific manner. Accordingly, a simple yet effective approach named PASE, i.e., partial multi-label learning via label-specific feature corrections, is proposed. Under a meta-learning framework, PASElearns to exert label-specific feature corrections so that potential ambiguity specific to each class label can be eliminated and the desired prediction model can be induced on these corrected instance features with the provided candidate labels. Comprehensive experiments on a wide range of synthetic and real-world data sets validate the effectiveness of the proposed approach.
In this paper, we consider a K-user interference channel where interference among the users is neither too strong nor too weak, a scenario that is relatively underexplored in the literature. We propose a novel deep le...
详细信息
Unmanned Aerial Vehicles(UAvs)as aerial base stations to provide communication services for ground users is a flexible and cost-effective paradigm in ***,dynamic resource allocation and multi-connectivity can be adopt...
详细信息
Unmanned Aerial Vehicles(UAvs)as aerial base stations to provide communication services for ground users is a flexible and cost-effective paradigm in ***,dynamic resource allocation and multi-connectivity can be adopted to further harness the potentials of UAVs in improving communication capacity,in such situations such that the interference among users becomes a pivotal disincentive requiring effective *** this end,we investigate the Joint UAV-User Association,Channel Allocation,and transmission Power Control(J-UACAPC)problem in a multi-connectivity-enabled UAV network with constrained backhaul links,where each UAV can determine the reusable channels and transmission power to serve the selected ground *** goal was to mitigate co-channel interference while maximizing long-term system *** problem was modeled as a cooperative stochastic game with hybrid discrete-continuous action space.A Multi-Agent Hybrid Deep Reinforcement Learning(MAHDRL)algorithm was proposed to address this *** simulation results demonstrated the effectiveness of the proposed algorithm and showed that it has a higher system utility than the baseline methods.
Sign language recognition is vital for enhancing communication accessibility among the Deaf and hard-of-hearing *** Japan,approximately 360,000 individualswith hearing and speech disabilities rely on Japanese Sign Lan...
详细信息
Sign language recognition is vital for enhancing communication accessibility among the Deaf and hard-of-hearing *** Japan,approximately 360,000 individualswith hearing and speech disabilities rely on Japanese Sign Language(JSL)for ***,existing JSL recognition systems have faced significant performance limitations due to inherent *** response to these challenges,we present a novel JSL recognition system that employs a strategic fusion approach,combining joint skeleton-based handcrafted features and pixel-based deep learning *** system incorporates two distinct streams:the first stream extracts crucial handcrafted features,emphasizing the capture of hand and body movements within JSL ***,a deep learning-based transfer learning stream captures hierarchical representations of JSL gestures in the second ***,we concatenated the critical information of the first stream and the hierarchy of the second stream features to produce the multiple levels of the fusion features,aiming to create a comprehensive representation of the JSL *** reducing the dimensionality of the feature,a feature selection approach and a kernel-based support vector machine(SVM)were used for the *** assess the effectiveness of our approach,we conducted extensive experiments on our Lab JSL dataset and a publicly available Arabic sign language(ArSL)*** results unequivocally demonstrate that our fusion approach significantly enhances JSL recognition accuracy and robustness compared to individual feature sets or traditional recognition methods.
Reducing a node’s power consumption is a difficult task for extending the network’s lifetime because the nodes are resource-constrained (i.e., limited battery power, processing capacity, storage, and non-rechargeabl...
详细信息
In the construction industry,to prevent accidents,non-destructive tests are necessary and *** impedance tomography is a new technology in non-invasive imaging in which the image of the inner part of conductive bodies ...
详细信息
In the construction industry,to prevent accidents,non-destructive tests are necessary and *** impedance tomography is a new technology in non-invasive imaging in which the image of the inner part of conductive bodies is reconstructed by the arrays of external electrodes that are connected on the periphery of the *** equipment is cheap,fast,and edge *** this imaging method,the image of electrical conductivity distribution(or its opposite;electrical impedance)of the internal parts of the target object is *** image reconstruction process is performed by injecting a precise electric current to the peripheral boundaries of the object,measuring the peripheral voltages induced from it and processing the collected *** an electrical impedance tomography system,the voltages measured in the peripheral boundaries have a non-linear equation with the electrical conductivity *** paper presents a cheap electrical Impedance Tomography(EIT)instrument for detecting impurities in the concrete.A voltage-controlled current source,a micro-controller,a set of multiplexers,a set of electrodes,and a personal computer constitute the structure of the *** conducted tests on concrete with impurities show that the designed EIT system can reveal impurities with a good accuracy in a reasonable time.
The Internet of Things(IoT)has taken the interconnected world by *** to their immense applicability,IoT devices are being scaled at exponential proportions ***,very little focus has been given to securing such *** the...
详细信息
The Internet of Things(IoT)has taken the interconnected world by *** to their immense applicability,IoT devices are being scaled at exponential proportions ***,very little focus has been given to securing such *** these devices are constrained in numerous aspects,it leaves network designers and administrators with no choice but to deploy them with minimal or no security at *** have seen distributed denial-ofservice attacks being raised using such devices during the infamous Mirai botnet attack in *** we propose a lightweight authentication protocol to provide proper access to such *** have considered several aspects while designing our authentication protocol,such as scalability,movement,user registration,device registration,*** define the architecture we used a three-layered model consisting of cloud,fog,and edge *** have also proposed several pre-existing cipher suites based on post-quantum cryptography for evaluation and *** also provide a fail-safe mechanism for a situation where an authenticating server might fail,and the deployed IoT devices can self-organize to keep providing services with no human *** find that our protocol works the fastest when using ring learning with *** prove the safety of our authentication protocol using the automated validation of Internet security protocols and applications *** conclusion,we propose a safe,hybrid,and fast authentication protocol for authenticating IoT devices in a fog computing environment.
In recent years,developed Intrusion Detection Systems(IDSs)perform a vital function in improving security and anomaly *** effectiveness of deep learning-based methods has been proven in extracting better features and ...
详细信息
In recent years,developed Intrusion Detection Systems(IDSs)perform a vital function in improving security and anomaly *** effectiveness of deep learning-based methods has been proven in extracting better features and more accurate classification than other *** this paper,a feature extraction with convolutional neural network on Internet of Things(IoT)called FECNNIoT is designed and implemented to better detect anomalies on the ***,a binary multi-objective enhance of the Gorilla troops optimizer called BMEGTO is developed for effective feature ***,the combination of FECNNIoT and BMEGTO and KNN algorithm-based classification technique has led to the presentation of a hybrid method called *** the next step,the proposed model is implemented on two benchmark data sets,NSL-KDD and TON-IoT and tested regarding the accuracy,precision,recall,and Fl-score *** proposed CNN-BMEGTO-KNN model has reached 99.99%and 99.86%accuracy on TON-IoT and NSL-KDD datasets,*** addition,the proposed BMEGTO method can identify about 27%and 25%of the effective features of the NSL-KDD and TON-IoT datasets,respectively.
The rapid advancement and proliferation of Cyber-Physical Systems (CPS) have led to an exponential increase in the volume of data generated continuously. Efficient classification of this streaming data is crucial for ...
详细信息
Advances in Internet of Things (IoT) wearable sensors and edge-artificial intelligence (Edge-AI) have enabled practical realizations of machine learning (ML)-enabled mobile sensing applications like human activity rec...
详细信息
暂无评论