A5101827231- Wireless Communication Security Techniques
- Blockchain Technology Applications and Security
- Smart Grid Security and Resilience
- Anomaly Detection Techniques and Applications
- Parallel Computing and Optimization Techniques
- Advanced Malware Detection Techniques
- IoT and Edge/Fog Computing
- Distributed and Parallel Computing Systems
- Energy Load and Power Forecasting
- Electric Vehicles and Infrastructure
- Sensorless Control of Electric Motors
- Smart Grid Energy Management
- Distributed Control Multi-Agent Systems
- Face and Expression Recognition
- Network Security and Intrusion Detection
- Domain Adaptation and Few-Shot Learning
- UAV Applications and Optimization
- Adversarial Robustness in Machine Learning
- Computer Graphics and Visualization Techniques
- Infrastructure Resilience and Vulnerability Analysis
- Power Line Inspection Robots
- Topic Modeling
- Advanced Graph Neural Networks
- 3D Shape Modeling and Analysis
- Smart Grid and Power Systems
University of California, Riverside
2024
University of Hong Kong
2024
National University of Defense Technology
2020-2023
University of Illinois Urbana-Champaign
2019-2023
Chongqing Normal University
2022
China Southern Power Grid (China)
2020-2022
Harbin Institute of Technology
2022
University of California System
2022
Jiangsu Provincial Key Laboratory of Network and Information Security
2021
Chinese University of Hong Kong
2018
We introduce phi-3-mini, a 3.8 billion parameter language model trained on 3.3 trillion tokens, whose overall performance, as measured by both academic benchmarks and internal testing, rivals that of models such Mixtral 8x7B GPT-3.5 (e.g., phi-3-mini achieves 69% MMLU 8.38 MT-bench), despite being small enough to be deployed phone. The innovation lies entirely in our dataset for training, scaled-up version the one used phi-2, composed heavily filtered web data synthetic data. is also further...
The ever-growing malware threats in the cyber spacecalls for techniques that are more effective than widely deployedsignature-based detection system. To counter large volumes ofmalware variants, machine learning have been appliedfor automated classification. Despite these efforts haveachieved a certain success, accuracy and efficiency stillremained inadequate to meet demand, especially when thesemachine used situation of multipleclass classification imbalanced training data. Against...
This paper investigates the destination image of Hong Kong UNESCO Global Geopark by comparing projected held park authority and perceived Mainland Chinese visitors. Through a qualitative categorization elements in official promotional materials, this study identifies categories projection, which concentrate on key characteristics geological features, interpretations education, as well accessibility proximity to urban areas. The survey-based approach reveals that visitors consider marine...
Resilience provides a new approach that system administrators can use in the design and analysis of engineering systems to enhance ability such withstand uncertain threats. In this article, an improved integrated metric is proposed for quantitative assessment resilience. The constructed form summation two capacities: absorptive restorative capacities. A weight coefficient assigned each capacity flexibility according various requirements stakeholders. addition, based on absolute time scale,...
Unbiased learning to rank (ULTR) studies the problem of mitigating various biases from implicit user feedback data such as clicks, and has been receiving considerable attention recently. A popular ULTR approach for real-world applications uses a two-tower architecture, where click modeling is factorized into relevance tower with regular input features, bias bias-relevant inputs position document. successful factorization will allow be exempt biases. In this work, we identify critical issue...
Through the analysis of wind turbines, photovoltaic systems, fuel cells and high frequency micro their models in power flow calculation were established this paper. Moreover, a novel algorithm based on back/forward sweep technology for radial distribution networks considering Distributed Generation (DG) is proposed. The develops reactive compensation method applying reactance matrix to handle all PV nodes studies (PFSs). A 33 bus test system used validate proposed method. simulation results...
MmWave MIMO communication is an emerging technique to enable the gigabit-per-second data rates for Internet of Things (IoT) networks. However, mmWave IoT networks face serious security spoofing attack threats. The lightweight channel-based physical layer authentication can detect attacks in networks, which has attracted great attentions. current schemes are mainly focused on sub-6 GHz microwave systems. In this work, we study efficient channel based by using sparse propriety channel. By...
With the widely use of cyber threat intelligence, influence security threats and attacks have been relieved controlled in a degree. More more users accepted conception intelligence are trying to routine protection. Then, how choose appropriate vendors services has become crucial issue. The present research evaluation mainly focused on one-sided contents approaches, which was lack comprehensiveness effectiveness. Aiming at this situation, we propose comprehensive architecture user perspective...
With the growing number of electric vehicles, integration large-scale disorderly charging loads into power grid will have some adverse impacts on safe and stable operation grid. As connection point between Internet grid, facilities face threat cyberattacks, which further threatens stability To illustrate potential impact cyberattacks to security, this paper proposes two attack scenarios corresponding strategies based law vehicles. Firstly, a spatial-temporal forecast model is established...
Thyroid nodule ultrasound image has serious noise, low contrast between different tissues, blurred boundary and irregular shape in malignant lesions. However, existing segmentation algorithms of thyroid often have problems coarse edge inaccurate small nodule. In order to segment nodules from images more precisely, this paper proposes a new deep learning algorithm, ASPP- UN et, achieve precise semantic images. the final phase encoding ASPP module is introduced, using void rate atrous...
Existing LLM training and inference frameworks struggle in boosting efficiency with sparsity while maintaining the integrity of context model architecture. Inspired by sharding concept database fact that attention parallelizes over heads on accelerators, we propose Sparsely-Sharded (S2) Attention, an algorithm allocates heterogeneous partitions for different to divide conquer. S2-Attention enforces each head only attend a partition contexts following strided pattern, full is preserved as...
Training and serving long-context large language models (LLMs) incurs substantial overhead. To address this, two critical steps are often required: a pretrained LLM typically undergoes separate stage for context length extension by training on data, followed architectural modifications to reduce the overhead of KV cache during serving. This paper argues that integrating with GPU-friendly reduction architecture not only reduces extension, but also achieves better performance. leads our...
How to control the process of continuously variable transmission is an important issue in electric system a tracked vehicle. It presented this paper that constant power brushless DC motor (BLDCM) controlled by feedback (observation) power, and it proved feasible on testing. Based above, range speed expanded winding change namely configuration change, rotational ranges corresponding forms are settled our study. Finally, some total performance simulation results showed analyzed, which lay...
Abstract With the advent of 5G era, number UAV (Unmanned Aerial Vehicle) has grown dramatically. Accordingly, need more computing and storage capabilities to complete a variety tasks (including some collaborative tasks). To meet requirements terminals, edge nodes devices be deployed. But problem trust between will lead series data security problems. The blockchain features, which contain detrust, immutable, decentralized, can solve this properly. Therefore, paper establishes an aviation...
Internet of Things (IoT) enable information transmission and sharing among massive IoT devices. However, the key establishment management in become more challenging due to low latency requirements resource constrained In this work, we propose a practical physical layer based secret scheme for MIMO (multiple-input-multiple-output) devices reduce communication delay caused by This is because proposed attachs with simultaneously. It achieved self-injection AN (SAN) tranmsission, which designed...
Abstract In this paper, particle swarm optimization (PSO) is introduced to carry out economic under the microgrid system, and dispatching problem of system [1] studied. Utilize role battery charging discharging power exchange large grids, with goal minimizing cost load supply, establishing constraints distributed sources [2]. A mathematical model established for optimal scheduling composed wind, solar storage batteries, PSO algorithm used solve model, optimized supply composition operating...
Abstract In recent years, electric vehicles (EVs) have gradually become the development trend of future automotive industry due to its advantages zero emission, low noise, and high efficiency. At this stage, a large number EV freely charging without guidance will bring many negative effects on operation urban traffic power distribution systems. Therefore, how correctly guide vehicle charge efficiently has an urgent problem be solved. Starting from economic factors, time user characteristics,...
Native English-Speaking Teachers (NESTs) have long been assumed as monolingual, and the translanguaging practices multilingual identity construction of NESTs under-explored. This study investigates how construct in online English teaching videos through translanguaging. Focusing on one NEST a popular Chinese social media platform, combination multimodal video analysis qualitative content was conducted. The findings reveal that breaks linguistic barriers, incorporating gestures, body...