A5044971195- Microfluidic and Capillary Electrophoresis Applications
- Data Quality and Management
- Machine Learning and Data Classification
- Advancements in Battery Materials
- Imbalanced Data Classification Techniques
- Lipid Membrane Structure and Behavior
- Scientific Computing and Data Management
- Protein Interaction Studies and Fluorescence Analysis
- Aluminum Alloys Composites Properties
- Natural Language Processing Techniques
- Advanced battery technologies research
- Mass Spectrometry Techniques and Applications
- Data Management and Algorithms
- Data Stream Mining Techniques
- Machine Learning and Algorithms
- Analytical Chemistry and Chromatography
- Mechanical Behavior of Composites
- Structural Behavior of Reinforced Concrete
- Web Data Mining and Analysis
- Advanced biosensing and bioanalysis techniques
- Aluminum Alloy Microstructure Properties
- Academic integrity and plagiarism
- Composite Structure Analysis and Optimization
- Analytical Chemistry and Sensors
- Electrowetting and Microfluidic Technologies
Beijing Institute of Technology
2008-2025
Ningbo University of Technology
2024
Wuhan University
2008
Northwestern Polytechnical University
2006-2008
Hubei University of Education
2008
National Taiwan University
2004
Supervised machine learning (ML) models trained on data with mislabeled instances often produce inaccurate results due to label errors. Traditional methods of detecting rely proximity, where an instance is considered if its inconsistent neighbors. However, it performs poorly, because does not always share the same ML-based instead utilize differentiate between and clean instances. these struggle achieve high accuracy, since may have already overfitted In this paper, we propose a novel...
An extended crack band model is presented to simulate the damaging behavior in fiber-reinforced composite laminates. The makes characteristic length of finite element and bandwidth independent each other, effectively relieving mesh dependency. It avoids problem big elements causing a snap-back at local constitutive level. dissipated energy per unit volume no longer affected by length, which solves excessive longitudinal fracture toughness. present was applied an open-hole tension test,...
While machine learning techniques, especially deep neural networks, have shown remarkable success in various applications, their performance is adversely affected by label errors training data. Acquiring high-quality annotated data both costly and time-consuming real-world scenarios, requiring extensive human annotation verification. Consequently, many industry-applied models are trained over containing substantial noise, significantly degrading the of these models.
Searching tables from poorly maintained data lakes has long been recognized as a formidable challenge in the realm of management. There are three pivotal tasks: keyword-based, joinable and unionable table search, which form backbone tasks that aim to make sense diverse datasets, such machine learning. In this demo, we propose LakeCompass, an end-to-end prototype system maintains abundant tabular data, supports all above search with high efficacy, well serves downstream ML modeling. To be...
Outlier detection is crucial for preventing financial fraud, network intrusions, and device failures. Users often expect systems to automatically summarize interpret outlier results reduce human effort convert outliers into actionable insights. However, existing methods fail effectively assist users in identifying the root causes of outliers, as they only pinpoint data attributes without considering same subspace may have different causes. To fill this gap, we propose STAIR, which learns...
Discovering tables from poorly maintained data lakes is a significant challenge in management. Two key tasks are identifying joinable and unionable tables, crucial for integration, analysis, machine learning. However, there's lack of comprehensive benchmark evaluating existing methods. To address this, we introduce LakeBench, large-scale table discovery benchmark. It evaluates effectiveness, efficiency, scalability join & union search With over 16 million real LakeBench 1,600X larger...
Massive spent Zn-MnO2 primary batteries have become a mounting problem to the environment and consume huge resources neutralize waste. This work proposes an effective recycling route, which converts MnO2 in LiMn2O4 (LMO) without any environmentally detrimental byproducts or energy-consuming process. The recycled LMO shows pure phase high-crystallinity properties, more importantly, when used as lithium-ion cathodes, exhibits impressive reversible capacity of 111.9 mAh g–1 excellent cycling...
A method for the investigation on interaction between bovine serum albumin (BSA) and liposome using capillary electrophoresis was developed. The oxidation index showed that liposomes after freeze-drying were more stable. results obtained from electrophoretic analysis of had no charge at pH 5.0 - 8.0. series suspension different concentrations with internal marker 0.8% dimethyl sulfoxide (DMSO) introduced as buffer 7.0. Along raised 0 to 2.4 mg/mL, it found effective mobility BSA changed...