A5107446421- Network Security and Intrusion Detection
- Topic Modeling
- Distributed and Parallel Computing Systems
- Advanced Malware Detection Techniques
- Global Financial Crisis and Policies
- Gait Recognition and Analysis
- Business Process Modeling and Analysis
- Model-Driven Software Engineering Techniques
- Cloud Computing and Resource Management
- Natural Language Processing Techniques
- Bioinformatics and Genomic Networks
- Advanced SAR Imaging Techniques
- Banking stability, regulation, efficiency
- Multimodal Machine Learning Applications
- Radar Systems and Signal Processing
- Language, Metaphor, and Cognition
- Aviation Industry Analysis and Trends
- Distributed systems and fault tolerance
- Imbalanced Data Classification Techniques
- Video Analysis and Summarization
- Advanced Software Engineering Methodologies
- Mental Health via Writing
- Advanced Thermodynamic Systems and Engines
- Creative Drama in Education
- Direction-of-Arrival Estimation Techniques
University of Rochester
2024
University of Wisconsin–Eau Claire
2024
University of Missouri
2020-2024
Ton Duc Thang University
2024
University of Alabama
2024
Lockheed Martin (United States)
2022-2024
Vietnam National University Ho Chi Minh City
2007-2023
University of Miami
2021-2022
Minnesota State University, Mankato
2021-2022
HEC Montréal
2022
We present DroidSafe, a static information flow analysis tool that reports potential leaks of sensitive in Android applications.DroidSafe combines comprehensive, accurate, and precise model the runtime with design decisions enable DroidSafe analyses to scale analyze this model.This combination is enabled by accurate stubs, technique enables effective code whose complete semantics lies outside scope Java, together can statically resolve communication targets identified dynamically constructed...
This paper describes transform coefficient coding in the draft international standard of High Efficiency Video Coding (HEVC) specification and driving motivations behind its design. Transform HEVC encompasses scanning patterns methods for last significant coefficient, significance map, levels, sign data. Special attention is paid to new coding, multilevel maps, high-throughput binarization, data hiding. Experimental results are provided evaluate performance HEVC.
Language prior plays an important role in the way humans detect and recognize text wild. Current scene recognition methods do use lexicons to improve performance, but their naive approach of casting output into a dictionary word based purely on edit distance has many limitations. In this paper, we present novel incorporate both training inference stage system. We generate list possible outcomes find one that is most compatible with visual appearance text. The proposed method leads robust...
Manual trial-and-error methods are employed for image parameter selection decisions in the processes that control cyber-enabled scientific instruments. Particularly materials manufacturing use cases, where analytics can be iterative, time-consuming and prone to errors, there is a need enhance existing by using agents featuring learning algorithms recommend parameters users. In this paper, we present an analysis focused on identifying optimal algorithm AI agent guiding real-time processing...
Container-based virtualization has recently been flourishing due to its ease of deployment and flexibility resource provisioning. Kubernetes, a well-known open source project, is an orchestration platform for containerized applications. Herein, applications can have multiple replicas provide high scalability availability, but some stateful service require leader be elected maintain consistency coordinate tasks among the replicas. In this article, Kubernetes architecture first analyzed by...
This paper studies communication patterns in mobile applications. Our analysis shows that 63% of the external made by top-popular free Android applications from Google Play has no effect on user-observable application functionality. To detect such covert an efficient manner, we propose a highly precise and scalable static technique: it achieves 93% precision 61% recall compared to empirically determined "ground truth", runs matter few minutes. Furthermore, according human evaluators, 42 out...
Being master regulators of gene expression, transcription factors (TFs) play important roles in determining plant growth, development and reproduction. To date, many TFs have been shown to positively mediate responses environmental stresses. In the current study, biological functions a stress-responsive NAC [NAM (No Apical Meristem), ATAF1/2 (Arabidopsis Transcription Activation Factor1/2), CUC2 (Cup-shaped Cotyledon2)]-TF encoding isolated from soybean (GmNAC019) relation drought tolerance...
Due to the extreme scarcity of customer failure data, it is challenging reliably screen out those rare defects within a high-dimensional input feature space formed by relevant parametric test measurements. In this paper, we study several unsupervised learning techniques based on six industrial datasets, and propose train more robust model self-labeling training data via set transformations. Using labeled multi-class classifier through supervised training. The goodness multiclass...
Objective: This study estimates the effects of generic competition, increased cost-sharing, and benefit practices on utilization spending for prescription drugs. Data Methods:We examined changes in Medicare price from 2007 to 2009 all drugs 28 therapeutic classes.The classes accounted 80% Part D included 6 protected with practically no competition.All variables were constructed measure each drug relative its class at a specific plan sponsor. Results:We estimated that shift toward had cut...
A framework for understanding the effect of various sources uncertainty in National Airspace System has been created and populated with data from past efforts to quantify such sources. This identifies specific agent functions NAS that are affected by these uncertainties indicates their interactions. In addition, effort produced a comprehensive library can be used future simulation efforts.
One important component of grid software infrastructure and parallel systems management is the Job Management System (JMS). With many JMSs available commercially in public domain, it difficult to choose most efficient JMS for a given computing environment. All previous comparisons had only conceptual character. In this paper, we present results first empirical study reported literature. Two popular commercial systems, LSF PBS Pro, were included our study. The has revealed strengths...
In heterogeneous networks, packet losses arise as a result of both congestion and random transmission errors. these the Transmission Control Protocol (TCP) performs poorly because it assumes all are caused by subsequently throttles its rate unnecessarily. This motivates need to discriminate between different types loss. this paper, we propose novel loss discrimination algorithm. Its design guided queueing analysis, our algorithm is based on unique definition customer, Lindley's Equation...
Vietnamese use of proverbs in everyday conversation, literature, and pedagogical exercises is well recognized. The AA. demonstrate that some provide the opportunity for intimates to offer psychologically sensitive insights a person manifesting problems. cited here an awareness conflictual affects can be held secret inside it therapeutic comment on them so they might dealt with openly consciously. Such reveals psychological mindedness often not revealed by other idioms communication among Vietnamese.
Knowledge Aided High Dynamic Range Space-Time Adaptive Processing (KA-HDR-STAP) approach enhances the STAP that uses Fast Slepian Transform (FST) to filter out more clutter energy than traditional STAP, Reduced Rank Eigen decomposition (RR) and KA-STAP approaches using typing prior STAP. relies on samples collected across sensor elements, space, time estimate interference with limitations due dynamic characteristics, spectrum spreading, limited separation between signals. KA-HDR-STAP a...
Incorporating linguistic knowledge can improve scene text recognition, but it is questionable whether the same holds for spotting, which typically involves detection and recognition. This paper proposes a method that leverages from large corpus to replace traditional one-hot encoding used in auto-regressive spotting recognition models. allows model capture relationship between characters word. Additionally, we introduce technique generate distributions align well with datasets, removing need...
The capability of intelligent models to extrapolate and comprehend changes in object states is a crucial yet demanding aspect AI research, particularly through the lens human interaction real-world settings. This task involves describing complex visual environments, identifying active objects, interpreting their as conveyed language. Traditional methods, which isolate captioning state change detection, offer limited view dynamic environments. Moreover, relying on small set symbolic words...
This project is part of DARPA Beyond Linear Processing (BLiP) that proposes integrating compressive sensing techniques, artificial intelligence, machine learning, and track-before-detect algorithms into a comprehensive processing string for radar systems. effort focuses on the control, interfaces, impact each algorithm has system's processing. The challenges result from designing to work together cohesively minimize signal loss been traditionally accepted within existing objective ability...
In the realm of physical activity and sports science, motion tracking emerges as an important tool for specialists aiming to analyze movement data, whether daily life activities or enhancing athletic performance. This study introduces innovative approach through development a fully embedded smart shoe sensor system, which employs inertial measurement unit (IMU) affixed sole, coupled with six pressure sensors, provide real-time calculation individual stride lengths along other vital...
Abstract This study aimed to simulate air pollution in Ba Ria - Vung Tau province (BR-VT), a significant transportation hub Vietnam and Southeast Asia. The FVM-TAPOM model system was employed create spatial distribution map of concentration including PM 2.5 , 10 NO 2 SO CO. had sufficient capability for simulating pollutant (R >0.6). For the average simulated 1-hour is highest during dry rainy seasons, reaching 68 μg/m 3 46 respectively. dust 90 62 56 (μg/m ) respectively, which lower...