A5028748926- UAV Applications and Optimization
- Anomaly Detection Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Distributed Control Multi-Agent Systems
- Adversarial Robustness in Machine Learning
- Vehicular Ad Hoc Networks (VANETs)
- Network Security and Intrusion Detection
- Advanced Wireless Communication Technologies
- Machine Learning and ELM
- Wireless Signal Modulation Classification
- Advanced Graph Neural Networks
- COVID-19 diagnosis using AI
- Multimodal Machine Learning Applications
- Explainable Artificial Intelligence (XAI)
- Neural Networks and Applications
- Advanced Neural Network Applications
- Blockchain Technology Applications and Security
- Coronary Interventions and Diagnostics
- Sparse and Compressive Sensing Techniques
- Speech Recognition and Synthesis
- AI in cancer detection
- Speech and Audio Processing
- Recommender Systems and Techniques
- Antiplatelet Therapy and Cardiovascular Diseases
- Protein Structure and Dynamics
Bowling Green State University
2021-2024
The University of Texas Health Science Center at Houston
2021-2024
Embry–Riddle Aeronautical University
2019-2021
Auburn University at Montgomery
2021
Shenzhen University
2021
Transfer learning (TL) has been successfully applied to many real-world problems that traditional machine (ML) cannot handle, such as image processing, speech recognition, and natural language processing (NLP). Commonly, TL tends address three main of learning: (1) insufficient labeled data, (2) incompatible computation power, (3) distribution mismatch. In general, can be organized into four categories: transductive learning, inductive unsupervised negative learning. Furthermore, each...
The Internet of Things (IoT) is becoming an indispensable part everyday life, enabling a variety emerging services and applications. However, the presence rogue IoT devices has exposed to untold risks with severe consequences. first step in securing detecting identifying legitimate ones. Conventional approaches use cryptographic mechanisms authenticate verify devices' identities. protocols are not available many systems. Meanwhile, these methods less effective when can be exploited or...
Medical image processing is one of the most important topics in Internet Things (IoMT). Recently, deep learning methods have carried out state-of-the-art performances on medical imaging tasks. In this paper, we propose a novel transfer framework for classification. Moreover, apply our method COVID-19 diagnosis with lung Computed Tomography (CT) images. However, well-labeled training data sets cannot be easily accessed due to disease's novelty and privacy policies. The proposed has two...
The Internet of Things (IoT) provides applications and services that would otherwise not be possible. However, the open nature IoT make it vulnerable to cybersecurity threats. Especially, identity spoofing attacks, where an adversary passively listens existing radio communications then mimic legitimate devices conduct malicious activities. Existing solutions employ cryptographic signatures verify trustworthiness received information. In prevalent IoT, secret keys for cryptography can...
Deep learning (DL) has been utilized pervasively in the Internet of Things (IoT). One typical application DL IoT is device identification from wireless signals, namely, noncryptographic (NDI). However, components NDI systems have to evolve adapt operational variations, such a paradigm termed as incremental (IL). Various IL algorithms proposed and many them require dedicated space store increasing amount historical data, therefore, they are not suitable for or mobile applications. Besides,...
Multi-agent unmanned aerial vehicle (UAV) teaming becomes an essential part in science mission, modern warfare surveillance, and disaster rescuing. This paper proposes a decentralized UAV swarm persistent monitoring strategy realizing continuous sensing coverage network service. A two-layer (high altitude low altitude) hierarchical structure is adopted the accurate object tracking area of interest (AOI). By introducing communication channel model its path planning, both centralized control...
The ubiquitous deployment of 5G New Radio (5G NR) accelerates the evolution in many fields. With enhancement NR, unmanned aerial vehicle (UAV) swarm networking can gain more flexibility, reliability, and elasticity to assist residents or workers finish missions high complexity risks remotely. beamforming NR improve accuracy flexibility connections between mobile devices. a throughput guaranteed UAV networking, remote deliver specific instructions with requirement. Further, reliable volume...
While deep-learning-enabled recommender systems demonstrate strong performance benchmarks, many struggle to adapt effectively in real-world environments due limited use of user-item relationship data and insufficient transparency recommendation generation. Traditional collaborative filtering approaches fail integrate multifaceted item attributes, although Factorization Machines account for item-specific details, they overlook broader relational patterns. Collaborative knowledge graph-based...
Big data analytics and mining have the potential to enable real-time decision making control in a range of Internet Things (IoT) application domains, such as Vehicles, Wings, Airport Things. The prediction toward air mobility, which is essential studies traffic management, has been challenging task due complex spatial temporal dependencies with highly nonlinear variational patterns. Existing works for only focus on either modeling static patterns individual flight or correlation, no limited...
Human activities and city routines follow patterns. Transfer learning can help achieve scalable solutions toward the realization of smart cities accounting for similarities between regions, domains, activities. In this study, we propose a transfer learning-based framework buildings to test hypothesis in energy-related problems. Our has two major components: network creation transferable predictive model. order create that groups sharing characteristics, evaluated strategies: novel clustering...
The rapid evolution of artificial intelligence (AI) in conjunction with recent updates dual antiplatelet therapy (DAPT) management guidelines emphasizes the necessity for innovative models to predict ischemic or bleeding events after drug-eluting stent implantation. Leveraging AI dynamic prediction has potential revolutionize risk stratification and provide personalized decision support DAPT management.
In this paper, we study a not well-investigated but important transfer learning problem termed Distant Domain Transfer Learning (DDTL). This topic is closely related to negative transfer. Unlike conventional problems which assume that the source domain and target are more or less similar each other, DDTL aims make efficient transfers even when domains tasks completely different. As an extreme example in image classification, there only sufficient amount of unlabeled images watches,...
The ubiquitous of 5G New Radio (5G NR) accelerates the massive implementations in many fields including swarm Unmanned Aircraft System (UAS) networking. ultra capacities NR can provide more sufficient networking services for UAS which enable to deploy complex and challenging scenarios achieve missions. However, conventional are mainly centralized or hierarchical is vulnerable dynamics deployment on a large scale. In this paper, we formulate cell wall communications heterogeneous with...
With the evolution of 5G cellular communication, beamforming is mature for implementation on a large scale. The development networking provides great opportunity swarm UAS. Concurrently, advantages UAS can provide immense improvement to advance industrial and residential implementations. However, nature antenna array constrains in limited space which rarely mentioned routing researches. In this paper, regarding constrained steering space, we proposed novel algorithm, Optimized Ad-hoc...
The ubiquitous deployment of 5G New Radio (5G NR) stimulates Unmanned Aircraft Systems (UAS) swarm networking to evolve achieve more imminent progress. heterogeneous collaboration between UAS enhances the complexity and efficiency mission complement that requires robustness, flexibility, sustainability throughput in networking. conventional approaches mainly are based on hierarchical architectures limited satisfy challenges with high dynamics a large scale. In this paper, we propose an...
In the past decades, information from all kinds of data has been on a rapid increase. With state-of-the-art performance, machine learning algorithms have beneficial for management. However, insufficient supervised training is still an adversity in many real-world applications. Therefore, transfer (TF) was proposed to address this issue. This article studies not well investigated but important TL problem termed cross-modality (CMTL). topic closely related distant domain (DDTL) and negative...
Recently, waste sorting has become more and important in our daily life. It plays an essential role the big picture of recycling, reducing environmental pollution significantly. Deep learning (DL) methods have been dominating field image classification successfully applied to tasks achieve state-of-art performance. However, most traditional DL require a massive amount annotated data for training phase. Unfortunately, there is only one small set sorting, TrashNet created by Standford. In...
The Internet of Things (IoT) is becoming an indispensable part everyday life, enabling a variety emerging services and applications. However, the presence rogue IoT devices has exposed to untold risks with severe consequences. first step in securing detecting identifying legitimate ones. Conventional approaches use cryptographic mechanisms authenticate verify devices' identities. protocols are not available many systems. Meanwhile, these methods less effective when can be exploited or...
Abnormal event detection with the lowest latency is an indispensable function for safety-critical systems, such as cyber defense systems. However, systems become increasingly complicated, conventional sequential methods less effective, especially when we need to define indicator metrics from complicated data manually. Although deep neural networks (DNNs) have been used handle heterogeneous data, theoretic assurability and explainability are still insufficient. This article provides a...
Deep Learning (DL) and Neural Networks (DNNs) are widely used in various domains. However, adversarial attacks can easily mislead a neural network lead to wrong decisions. Defense mechanisms highly preferred safety- critical applications. In this paper, firstly, we use the gradient class activation map (GradCAM) analyze behavior deviation of VGG-16 when its inputs mixed with perturbation or Gaussian noise. particular, our method locate vulnerable layers that sensitive We also show be detect...