A5100741632- Natural Language Processing Techniques
- Topic Modeling
- Educational Technology and Pedagogy
- Phonocardiography and Auscultation Techniques
- Advanced Technologies in Various Fields
- Complex Network Analysis Techniques
- Cardiovascular Disease and Adiposity
- Advanced Text Analysis Techniques
- Recommender Systems and Techniques
- Text and Document Classification Technologies
- Machine Learning and Algorithms
- Algorithms and Data Compression
- Advanced Steganography and Watermarking Techniques
- Chaos-based Image/Signal Encryption
- Internet Traffic Analysis and Secure E-voting
- Authorship Attribution and Profiling
- Computability, Logic, AI Algorithms
Xinjiang Medical University
2025
University of Massachusetts Lowell
2020-2024
Zhejiang Gongshang University
2019
Zhejiang Technical Institute of Economics
2019
Beijing Anzhen Hospital
2012
Capital Medical University
2012
To investigate the mechanisms by which Porphyromonas gingivalis (P. gingivalis) promotes malignant progression of oral squamous cell carcinoma (OSCC) through recruitment chemokine receptor 6-positive (CCR6+) regulatory T cells (Treg) in tumor microenvironment (TME). The Cancer Genome Atlas (TCGA) database was used to analyze correlation between ligand 20 (CCL20), CCR6, and Treg. Treg enrichment index expression levels interleukin (IL)-10 necrosis factor β1 (TGF-β1) were assessed high CCR6...
Chinese characters are a unique blend of language and art, featuring diverse artistic styles. Mastering these styles requires extensive practice limits public participation. To encourage broader participation, we developed real-time, interactive tool that supports multiple character art This uses diffusion model several LoRA models to capture the diversity art. It generates personalized, visually striking artworks in real-time by utilizing handwritten input, allowing users adjust various...
Generative text steganography has received considerable attention in the covert communication community for benefit of sending secret messages without need to modify carriers. Existing methods typically choose next word when generating a stego-text based on conditional probability encoding candidates, which may lead inadequate words underlying message. How generate semantically controllable with high capacity secure embedding message is main challenge. We address this challenge by proposing...
Existing tools to detect text generated by a large language model (LLM) have met with certain success, but their performance can drop when dealing texts in new domains. To tackle this issue, we train ranking classifier called RoBERTa-Ranker, modified version of RoBERTa, as baseline using dataset constructed that includes wider variety written humans and various LLMs. We then present method fine-tune RoBERTa-Ranker requires only small amount labeled data domain. Experiments show fine-tuned...
To explore the clinical application of spatio-temporal image correlation (STIC) in fetal echocardiography diagnosis congenital heart diseases. According to American Association Echocardiography guidelines and standards for performance echocardiogram, examined order 1286 cases, Checking STIC technology during cardiac volume data acquisition saving post-processing offline analysis, results malformations is analyzed contrast with after birth echocardiography, surgical treatment pathology...
Recommender systems are widely used to provide users with appropriate items and have emerged in response the problem of information overload. The measuring method user relationship social networks is core personalised recommendation. To a certain extent, quality users' determines accuracy recommendation results. In this paper, we design new which considers direct indirect relationships among users. experimental results on two famous microblog data sets show that presented algorithm high performance.