A5100407779- Topic Modeling
- Aquatic Ecosystems and Phytoplankton Dynamics
- Advanced Computational Techniques and Applications
- Sentiment Analysis and Opinion Mining
- Marine and coastal ecosystems
- Anomaly Detection Techniques and Applications
- Advanced Text Analysis Techniques
- Web Data Mining and Analysis
- Air Traffic Management and Optimization
- Forecasting Techniques and Applications
- Semantic Web and Ontologies
- Natural Language Processing Techniques
- Neural Networks and Applications
- Text and Document Classification Technologies
- Evaluation Methods in Various Fields
- Advanced Polymer Synthesis and Characterization
- Sleep and Wakefulness Research
- Speech and dialogue systems
- Advanced Software Engineering Methodologies
- Image and Object Detection Techniques
- Educational Technology and Pedagogy
- Engineering Education and Technology
- Biocrusts and Microbial Ecology
- Advanced Graph Neural Networks
- Software Engineering Research
Nanjing Institute of Geography and Limnology
2023-2025
Chinese Academy of Sciences
2023-2025
University of Chinese Academy of Sciences
2023-2025
Mitre (United States)
2020-2024
North China University of Science and Technology
2022-2024
State Key Laboratory of Pollution Control and Resource Reuse
2023
Tongji University
2023
Guizhou Provincial People's Hospital
2023
Wenzhou Vocational College of Science and Technology
2022
Emory University
2020
Positively charged micelles based on a triblock copolymer demonstrate enhanced corneal penetration Jingguo Li, Zhanrong Tianyang Zhou, Junjie Zhang, Huiyun Xia, Heng Jijun He, Siyu Liya Wang Henan Eye Institute, Hospital, Provincial People’s Hospital and Zhengzhou University Zhengzhou, People’s Republic of China Purpose: The cornea is main barrier to drug after topical application. aim this study was evaluate the abilities generated from positively penetrate Methods:...
The emotion-cause pair extraction (ECPE) task is to simultaneously extract emotions and causes as pairs (EC-pairs) from documents, which important for natural language processing. Previous research tackled this via a two-step approach, first predicts separately the emotion cause clauses, then them up by using binary classifier. However, such approach may suffer possible propagation of errors, it neglects interaction between causes. In article, an end-to-end double-graph method with...
ABSTRACT Exploring the response of natural community stability to anthropogenic environmental changes, such as eutrophication, is an important topic in current ecological research. Eutrophication directly affects species dynamics, abundance and succession phytoplankton communities, potentially leading shifts processes these communities over multiple years. However, it remains unclear how annual monthly dynamics shift along eutrophication gradients maintain stability. We conducted 8‐year...
Emotion-Cause Pair Extraction (ECPE) aims to jointly extract emotion clauses and the corresponding cause from a document, which is important for user evaluation or public opinion analysis. Existing research addresses ECPE task through two-step an end-to-end approach. Although previous work shows promising performances, they suffer two limitations: 1) fail take full advantage of type information, has advantages modelling dependencies between semantic perspective; 2) ignored interaction local...
A neural network based on a word or character embedding is mainstream model framework in text sentiment analysis and has achieved good results. However, there lack of learning about POS-Tagging Sequence-Tagging. In this research, we propose multifeature data-augmentation (M-DA) with multiple-input single-output structure to overcome problem Chinese analysis. First, paper sequentially obtains various sequences text, including sequence, pos char char_pos char_4tag use the construct new...
Learning effective sentence representations is crucial for many Natural Language Processing (NLP) tasks, including semantic search, textual similarity (STS), and clustering. While multiple transformer models have been developed embedding learning, these may not perform optimally when dealing with specialized domains like aviation, which has unique characteristics such as technical jargon, abbreviations, unconventional grammar. Furthermore, the absence of labeled datasets makes it difficult...
The detection and segmentation of neoplasms are an important part radiotherapy treatment planning, monitoring disease progression, predicting patient outcome. In the brain, functional magnetic resonance imaging (MRI) like dynamic susceptibility contrast enhanced (DSC) or T1-weighted (DCE) perfusion MRI tools for diagnosis. However, manual contouring these tedious, expensive, time-consuming, contains inter-observer variability. this work, we propose to use a 3D Mask R-CNN method automatically...
With the rapid development of text categorization technology, there are still some problems, such as low classification efficiency, accuracy and incomplete extraction features, in case large amount data too many categorized attributes. In this paper, a hybrid model CNN (Convolutional Neural Network) BiLSTM (Bidirectional Long-term Short-term Memory combined with Attention (Attention Mechanism) is used to classify process long data. extracts feature information from text, then uses extract...
Objectives: This study involved a multi-omics analysis of glioblastoma (GBM) samples to elaborate the potential mechanism drug treatment. Methods: The GBM cells treated with or without orexin A were acquired from sequencing analysis. Differentially expressed genes/proteins/metabolites (DEGs/ DEPs/ DEMs) screened. Next, combination analyses conducted investigate common pathways and correlations between two groups. Lastly, transcriptome-proteome-metabolome association was carried out determine...
This paper proposes a novel approach based on review mining to analyze product usability. The massive online customer reviews products is used as source data. Opinion technology adopted translate the unstructured into structured feature-opinion pairs. Then, factor analysis technique utilized extract pairs related usability for evaluation. Finally, case study presented prove effectiveness of proposed approach.
Product usability is an evaluation indicator of customer satisfaction with product features and functions. Questionnaire survey traditional methods widely used in feature analysis. In fact, it labour intensive time consuming. Nowadays, the rapid expansion web, more customers tend to express their opinions on products through it. As a result, number reviews certain grows rapidly. This paper proposes novel method based web semantic mining Chinese analyse usability. First, useful information...
The advent of ChatGPT and GPT-4 has captivated the world with large language models (LLMs), demonstrating exceptional performance in question-answering, summarization, content generation. aviation industry is characterized by an abundance complex, unstructured text data, replete technical jargon specialized terminology. Moreover, labeled data for model building are scarce this domain, resulting low usage data. emergence LLMs presents opportunity to transform situation, but there a lack...
Due to people’s increasing dependence on software, the emergence of software defects will lead serious consequences. And essential cause is complexity software. The premise reducing understand topology ensure quality. refers connection between internal elements and it has become an important factor affecting quality In this paper, we use complex network theory as a tool analyze topology. Firstly, extract structural information from source code system abstract extracted with theory. Secondly,...
Lots of clustering algorithms have been developed, and in most them some parameters should be determined by hand. However, it is very difficult to determine manually without any prior domain knowledge. To solve this problem, a novel similarity based algorithm was presented. It aimed at avoiding instructional hand, the same time, improving efficiency clustering. By introducing fuzzy Adaptive Resonance Theory (ART) pre-train criterion, parameter dynamically. The new analyzed applied later...
Abstract Event subject extraction was to extract subjects of specific event types. For the traditional BiLSTM network, threshold is complicated, required parameters are many, and time cost high. This paper oriented financial field proposes a method introducing multi-head attention mechanism based on BIGRU network subjects. First, text vectorized, then word vector obtained input into learn context features, introduce depth feature values text. Finally, comparative experiment conducted data...
View Video Presentation: https://doi.org/10.2514/6.2021-2388.vid The current practice of manually processing features for high-dimensional and heterogeneous aviation data is labor-intensive, does not scale well to new problems, prone information loss, affecting the effectiveness maintainability machine learning (ML) procedures. This research explored an unsupervised method, autoencoder, extract effective problems. study variants autoencoders with aim forcing learned representations input...