A5029797380- Topic Modeling
- Crystallization and Solubility Studies
- X-ray Diffraction in Crystallography
- Advanced Text Analysis Techniques
- Anomaly Detection Techniques and Applications
- Advanced Neural Network Applications
- Text and Document Classification Technologies
- Advanced Image and Video Retrieval Techniques
- Building Energy and Comfort Optimization
- Traffic Prediction and Management Techniques
- Power Systems and Renewable Energy
- Lung Cancer Treatments and Mutations
- Sustainable Building Design and Assessment
- Fibroblast Growth Factor Research
- Transportation Planning and Optimization
- Adaptive Dynamic Programming Control
- Computational and Text Analysis Methods
- Machine Learning and ELM
- Noise Effects and Management
- Multimodal Machine Learning Applications
- Explainable Artificial Intelligence (XAI)
- Cancer Diagnosis and Treatment
- Epigenetics and DNA Methylation
- Complex Network Analysis Techniques
- Visual Attention and Saliency Detection
University of Science and Technology of China
2024
Wuhan University
2024
Shenyang Sport University
2022
Dalian University of Technology
2019
Tongji University
2018
Zhejiang University
2013-2017
Zhejiang University of Science and Technology
2017
Abstract Background Bone or brain metastases may develop in 20–40% of individuals with late‐stage non‐small‐cell lung cancer (NSCLC), resulting a median overall survival only 4–6 months. However, the primary tissue's distinctions between bone, and intrapulmonary NSCLC at single‐cell level have not been underexplored. Methods We conducted comprehensive analysis 14 tissue biopsy samples obtained from treatment‐naïve advanced patients bone ( n = 4), 6) 4) metastasis using sequencing originating...
It is observed that distinct words in a given document have either strong or weak ability delivering facts (i.e., the objective sense) expressing opinions subjective depending on topics they associate with. Motivated by intuitive assumption different varying degree of discriminative power sense with respect to their assigned topics, model named as identified objective-subjective latent Dirichlet allocation (LDA) (iosLDA) proposed this paper. In iosLDA model, simple Pólya urn adopted...
Many of the words in a given document either deliver facts (objective) or express opinions (subjective), respectively, depending on topics they are involved in. For example, bunch documents, word “bug” assigned to topic “order Hemiptera” apparently remarks one object (i.e., kind insects), while same “software” probably conveys negative opinion. Motivated by intuitive assumption that different have varying degrees discriminative power delivering objective sense subjective with respect their...
The ever-increasing user-generated contents in social media and other web services make it highly desirable to discover opinions of users on all kinds topics. Motivated by the assumption that individual word paragraph documents will deliver fine-grained (e.g., "laudatory", "annoyed" or "boring") coarse-grained positive, negative neutral) sentiments about certain topics respectively, this paper focuses a deeper thematic level jointly disentangle towards terms sentiment analysis, named as LDA...
Benefiting from the progress of power electronics technology, distributed generation technology is developing rapidly. Since micro grids cannot rely on traditional multi-time scale control strategies to ensure high-quality frequency stability and economic dispatch in same time scale, this paper proposes an extreme dynamic programming algorithm. The proposed algorithm takes adaptive as framework, learning machine a kernel evaluation module, model implementation module new prediction module....
Nowadays, a considerably large number of documents are available over many online news sites (e.g., CNN and NYT). Therefore, the utilization these documents, for example, discovery burst topic its evolution, is significant challenge. In this paper, novel model, called intermittent Evolution LDA (iELDA) proposed. iELDA, time-evolving divided into small epochs. iELDA utilizes detected global topics as priors to guide detection an emerging keep track evolution different As natural extension...
While image captioning has been extensively studied, the problem of generating narrative descriptions for photo streams still remains under explored. Photo stream is more challenging due to large visual variance, complicated object context, and sentence-to-sentence coherence in ordered collection photos. To deal with these challenges, we propose a novel deep contextual attention network (CAN) narratively describe by jointly exploring rich context among attended regions sentences. The...
The recent rise of EEG-based end-to-end deep learning models presents a significant challenge in elucidating how these process raw EEG signals and generate predictions the frequency domain. This limits transparency credibility models, hindering their application security-sensitive areas. To address this issue, we propose mask perturbation method to explain behavior Considering characteristics data, introduce target alignment loss mitigate out-of-distribution problem associated with...
Segments, such as sentence boundaries in texts or annotated regions images, can be considered useful structural constraints (i.e., priors) for unsupervised topic modeling. However, some segment units (e.g., words visual images) inside a given may irrelevant to the of this due their characteristics. This paper proposes model called πLDA, which introduces latent variable π into LDA, traditional model, capture characteristic each unit. That is say, πLDA conducted determine whether unit assigned...
The research purpose is to study the standardization and scientizing of physical training actions. Stacking denoising auto encoder (SDAE), a BiLSTM deep network model (SDAL-DNM) (a kind action model), an unsupervised transfer are used deeply problem training. Initially, discrimination adopted here combination stacked noise reduction self-encoder bidirectional depth model. Then, this can collect data for five actions in further analyze importance Afterward, SDAL-DNM implemented fully...
With the rapid development of intelligent transportation system applications, a tremendous amount multi-view video data has emerged to enhance vehicle perception. However, performing analytics efficiently by exploiting spatial-temporal redundancy from remains challenging. Accordingly, we propose novel traffic-related framework named CEVAS achieve efficient object detection using data. Briefly, fine-grained input filtering policy is introduced produce reasonable region interest captured...
Multi-agent collaborative perception as a potential application for vehicle-to-everything communication could significantly improve the performance of autonomous vehicles over single-agent perception. However, several challenges remain in achieving pragmatic information sharing this emerging research. In paper, we propose SCOPE, novel framework that aggregates spatio-temporal awareness characteristics across on-road agents an end-to-end manner. Specifically, SCOPE has three distinct...
In terms of emergency solutions to sudden occasions occurred in the railway transportation, multi-intelligent model is introduced passengers, vehicles, and terminals so as make analysis predictions on passenger volumes railway. For this reason, volume matching proposed vehicle organization. The paper conducts visual simulation toward under ArcGIS software makes evaluations effect metrics, including mean delay times, occupancy rates total passengers. Finally, provides corresponding technology...