A5101083689- Solar Radiation and Photovoltaics
- Reinforcement Learning in Robotics
- Microgrid Control and Optimization
- Topic Modeling
- Text and Document Classification Technologies
- Photovoltaic System Optimization Techniques
- Smart Grid Energy Management
- Electricity Theft Detection Techniques
- Data Stream Mining Techniques
- Evolutionary Algorithms and Applications
- Cutaneous Melanoma Detection and Management
- Optimal Power Flow Distribution
- Big Data and Business Intelligence
- Machine Learning in Bioinformatics
- Spectroscopy and Chemometric Analyses
- Remote Sensing and Land Use
- Energy Load and Power Forecasting
- Spectroscopy Techniques in Biomedical and Chemical Research
- Expert finding and Q&A systems
- Natural Language Processing Techniques
- Traffic control and management
- Image Retrieval and Classification Techniques
- Power Systems and Renewable Energy
- Adaptive Dynamic Programming Control
- Image Enhancement Techniques
Florida Atlantic University
2022-2025
Xiamen University
2020
Shenzhen Research Institute of Big Data
2018-2019
Chinese University of Hong Kong, Shenzhen
2019
Learning control in environments with uncertainties and perturbations remains a challenging issue the field of artificial intelligence. Though conventional imitation learning (IL) inverse reinforcement (IRL) methods have made some progress handling perturbations, repeatability resilience are somehow limited. To alleviate this issue, we propose multi-virtual-agent IRL (MVIRL) method to produce stable policies. Specifically, design multiple virtual agents interacting pertinent environments....
Text classification is a fundamental task in Natural Language Processing (NLP), and the advent of Large Models (LLMs) has revolutionized field. This paper introduces Smart Expert System, novel approach that leverages LLMs as text classifiers. The system simplifies traditional workflow, eliminating need for extensive preprocessing domain expertise. performance several LLMs, machine learning (ML) algorithms, neural network (NN) based structures evaluated on four datasets. Results demonstrate...
Using Raman spectroscopy (RS) signals for skin cancer tissue classification has recently drawn significant attention, because of its non-invasive optical technique nature using molecular structures and conformations within biological diagnosis. In reality, RS are noisy unstable training machine learning models. The scarcity samples also makes it challenging to learn reliable deep-learning networks clinical usages. this paper, we advocate a Transfer Contrasting Learning Paradigm (TCLP)...
With the increasing popularity of integrating solar energy into power system, prediction has recently attracted much interest, where movement clouds a crucial impact on irradiance and is major cause rapid, violent irregular fluctuations production. Meanwhile, it necessary for to capture these several minutes ahead in order facilitate scheduling operations that maintain system stability. Considering such importance cloud movement, sky images provided by all-sky cameras consist important data...
To address the serious challenges to traditional power systems caused by high variability of solar production, especially ramp events, this paper is aimed at predicting radiation accurately a very-short-term scale, i.e. several minutes ahead. Quite different from methods, our designed system conducts time series prediction based upon meteorological data and sky images' features with Long Short-Term Memory (LSTM) network learn long-term dependency. Experiments have been conducted on feature...
The efficacy and ethical integrity of large language models (LLMs) are profoundly influenced by the diversity quality their training datasets. However, global landscape data accessibility presents significant challenges, particularly in regions with stringent privacy laws or limited open-source information. This paper examines multifaceted challenges associated acquiring high-quality for LLMs, focusing on scarcity, bias, low-quality content across various linguistic contexts. We highlight...
Increasing popularity of integrating distributed energy resources (DERs) into the power system brings a challenge to optimize dispatch policy for microgrid scheduling. The reinforcement learning methods suffer from long-time problem with empirical assumption reward function system. Although traditional inverse (IRL) approaches can solve this some extent, they encounter limitation extensive computations state visitation frequency in large and continuous space. To alleviate limitation, we...
Driven by a plethora of real machine learning applications, there have been many attempts at improving the performance classifier applied to imbalanced dataset. In this paper we propose maximum entropy (MEM) based hybrid algorithm handle binary classification problems with high imbalance ratios and large numbers features in datasets. At training stage, combine an efficient MEM SMOTE build batch manner. application different-cost strategy is incorporated into problem online Experiments are...
Generalization problem of reinforcement learning is crucial especially for dynamic environments. Conventional methods solve the problems with some ideal assumptions and are difficult to be applied in environments directly. In this paper, we propose a new multi-virtual- agent (MVARL) approach predator-prey grid game. The designed method can find optimal solution even when predator moves. Specifically, design virtual agents interact simulated changing parallel instead using actual agents....