A5101889270- Natural Language Processing Techniques
- Topic Modeling
- Advanced Text Analysis Techniques
- Human Pose and Action Recognition
- Multimodal Machine Learning Applications
- Video Analysis and Summarization
- Data Mining Algorithms and Applications
- Visual Attention and Saliency Detection
- Text and Document Classification Technologies
- Advanced Image and Video Retrieval Techniques
- Semantic Web and Ontologies
- Wikis in Education and Collaboration
- Music and Audio Processing
- Information Retrieval and Search Behavior
- Advanced Clustering Algorithms Research
- Sentiment Analysis and Opinion Mining
- Gait Recognition and Analysis
- Music Technology and Sound Studies
- Neural Networks and Applications
- Knowledge Management and Sharing
- Wireless Signal Modulation Classification
- Robot Manipulation and Learning
- Economic Growth and Development
- Optimization and Variational Analysis
- Sparse and Compressive Sensing Techniques
Civil Aviation University of China
2022-2024
Harbin Engineering University
2024
Southwestern University of Finance and Economics
2024
Massey University
2024
Zhejiang University
2023
Shandong University
2022
Shanghai University of Traditional Chinese Medicine
2021
Lanzhou Jiaotong University
2021
Hebei University of Technology
2021
Kuaishou (China)
2019
Recently, skeleton-based human action recognition has attracted a lot of research attention in the field computer vision. Graph convolutional networks (GCNs), which model body skeletons as spatial-temporal graphs, have shown excellent results. However, existing methods only focus on local physical connection between joints, and ignore non-physical dependencies among joints. To address this issue, we propose hypergraph neural network (Hyper-GNN) to capture both information high-order for...
Keyphrases can concisely describe the high-level topics discussed in a document that usually possesses hierarchical topic structures. Thus, it is crucial to understand structures and employ guide keyphrase identification. However, integrating information into deep generation model unexplored. In this paper, we focus on how effectively exploit improve performance (HTKG). Specifically, propose novel topic-guided variational neural sequence method for generation, which consists of two major...
Current video-based scene graph generation (VidSGG) methods have been found to perform poorly in predicting predicates that are less represented due the inherently biased distribution of training data. In this paper, we take a closer look at inherent characteristics and identify most visual relations (e.g. <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">sit_above</monospace> ) involve both actional pattern (...
The purpose of this study is to examine the impact various factors on level non-performing loans (NPL) and, determine moderating role efficiency relationship between different and NPL in China. current addressed four important relation NPL. These include; return assets (ROA), equity (ROE), economic sustainability political instability index. Furthermore, these Secondary data used consider empirical results. related ROA, ROE, index collected from sources. Consistent with literature, we found...
Automatic singing evaluation without reference melody has long been a difficult problem. This paper aims to pilot novel data driven approach tackle this artistic We constructed large scale dataset and designed an innovative Bi-Dense neural network which can address task efficiently. Though the is quite subjective depends lot on listeners' preferences, we showed that specific group consistency evaluations, it possible train model learn preferences of group. In paper, amount clips...
Video-based scene graph generation (VidSGG) is an approach that aims to represent video content in a dynamic by identifying visual entities and their relationships. Due the inherently biased distribution missing annotations training data, current VidSGG methods have been found perform poorly on less-represented predicates. In this paper, we propose explicit solution address under-explored issue supplementing predicates should be included ground-truth annotations. Dubbed Trico, our method...
In this paper, we present a novel frequent generalized pattern mining algorithm, called GP-Close, for associations from RDF metadata. To solve the over-generalization problem encountered by existing methods, GP-Close employs notion of emphgeneralization closure systematic reduction.
In recent years, neural networks have been widely used in natural language processing, especially sentence similarity modeling. Most of the previous studies focused on current sentence, ignoring commonsense knowledge related to task Commonsense can be remarkably useful for understanding semantics sentences. CK-Encoder, which effectively acquire improve performance modeling, is proposed this paper. Specifically, model first generates a graph input and calculates by using convolution network....
This paper firstly introduces the definition and types of wavelet transform. Then we introduce Mallat decomposition reconstruction algorithm in detail. Secondly, after comparing advantages disadvantages several Better Method, choose thresholding make improvements that. At last, compare SNR signal de-noising by matlab simulation analysis to prove superiority algorithm.
An instant question answering system ProMe is proposed in this paper which sets automated as prototype and combines with the search engines messenger. For any queried by a user, looks for answers flow of three methods. The first one to match questions FAQ database TF-IDF (Term Frequency-Inverse Document Frequency) algorithm. If still not found, submit it some get extracting information from web pages returned engine. answer found or satisfied after evaluation, will be sent other users who...
KG-to-text generation models aim to generate a natural language text with one or more sentences from input knowledge graphs (KGs). Recent use either entity-level word-level as data texts. The Entity-level graph provides the relations between entities but neglects words within single entity and different entities. In contrast, can better consider rich semantic among fail capture This paper proposes novel graph-to-sequence model that encodes both KGs enhance node representations investigates...
Primal-dual hybrid gradient (PDHG) is a first-order method for saddle-point problems and convex programming introduced by Chambolle Pock. Recently, Applegate et al.\ analyzed the behavior of PDHG when applied to an infeasible or unbounded instance linear programming, in particular, showed that able diagnose these conditions. Their analysis hinges on notion infimal displacement vector closure range mapping splitting operator encodes algorithm. In this paper, we develop novel formula using...