A5100773559- Traffic Prediction and Management Techniques
- Human Mobility and Location-Based Analysis
- Recommender Systems and Techniques
- Transportation Planning and Optimization
- Advanced Graph Neural Networks
- Topic Modeling
- Network Security and Intrusion Detection
- Cloud Computing and Resource Management
- Machine Learning and Data Classification
- Anomaly Detection Techniques and Applications
- Data Management and Algorithms
- Advanced Data and IoT Technologies
- Graph Theory and Algorithms
- Caching and Content Delivery
- Network Traffic and Congestion Control
- Complex Network Analysis Techniques
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- Text and Document Classification Technologies
- Data Mining Algorithms and Applications
- Human Pose and Action Recognition
- Service-Oriented Architecture and Web Services
- Advanced Bandit Algorithms Research
- Mental Health via Writing
- Face and Expression Recognition
China Mobile (China)
2010-2025
Huaihua University
2025
Xi'an Jiaotong University
2023
Beijing University of Posts and Telecommunications
2023
Xiamen University of Technology
2023
Tsinghua University
2022
Alibaba Group (China)
2019
Guangdong Polytechnic of Science and Technology
2016
Beijing Jiaotong University
2014
Dalian University of Technology
2008-2010
In recent years, vision transformers have been introduced into face recognition and analysis achieved performance breakthroughs. However, most previous methods generally train a single model or an ensemble of models to perform the desired task, which ignores synergy among different tasks fails achieve improved prediction accuracy, increased data efficiency, reduced training time. This paper presents multi-purpose algorithm for simultaneous recognition, facial expression age estimation,...
Topology impacts important network performance metrics, including link utilization, throughput and latency, is of central importance to operators. However, due the combinatorial nature topology, it extremely difficult obtain an optimal solution, especially since topology planning in networks also often comes with management-specific constraints. As a result, local optimization hand-tuned heuristic methods from human experts adopted practice. Yet, cannot cover global design space while taking...
With the rapid growth of data and computing power, deep learning based approaches have become main solution for many artificial intelligence problems such as image classification, speech recognition computer vision. Several excellent (DL) frameworks including Tensorflow, MxNet PyTorch been made open-sourced, further accelerating advance community. However, existing DL are not designed applications involving high-dimensional sparse data, which exists widely in successful online businesses...
Understanding mobile data traffic and forecasting future trend is beneficial to wireless carriers service providers who need perform resource allocation energy saving management. However, predicting accurately at large-scale fine-granularity particularly challenging due the following two factors: spatial correlations between network units (i.e., a cell tower or an access point) introduced by user arbitrary movements, time-evolving nature of movements which frequently changes with time. In...
ABSTRACT Making medication prescriptions in response to the patient's diagnosis is a challenging task. The number of pharmaceutical companies, their inventory medicines, and recommended dosage confront doctor with well-known problem information cognitive overload. To assist medical practitioner making informed decisions regarding prescription patient, researchers have exploited electronic health records (EHRs) automatically recommending medication. In recent years, recommendation using EHRs...
In this study, we propose a method named Semantic Graph Neural Network (SGNN) to address the challenging task of email classification. This converts classification problem into graph by projecting and applying SGNN model for The features are generated from semantic graph; hence, there is no need embedding words numerical vector representation. performance tested on different public datasets. Experiments in dataset show that presented achieves high accuracy test against few better than...
Sequential recommendation aims at predicting the next item that user may be interested in given historical interaction sequence. Typical neural models derive a single history embedding to represent user's interests. Moving one step forward, recent studies point out multiple sequence embeddings can help better capture multi-faceted However, when ranking candidate items, these methods usually adopt greedy inference strategy. This approach uses best matching interest for each calculate score,...
Graphs are widely used to represent the relations among entities. When one owns complete data, an entire graph can be easily built, therefore performing analysis on is straightforward. However, in many scenarios, it impractical centralize data due privacy concerns. An organization or party only keeps a part of whole i.e., isolated from different parties. Recently, Federated Learning (FL) has been proposed solve isolation issue, mainly for Euclidean data. It still challenge apply FL because...
As psychological diseases become more prevalent and are identified as the leading cause of acquired disability, it is essential to assist people in improving their mental health. Digital therapeutics (DTx) has been widely studied treat with advantage cost savings. Among techniques DTx, a conversational agent can interact patients through natural language dialog most promising one. However, agents' ability accurately show emotional support (ES) limits role DTx solutions, especially health...
Rural distribution networks have complex structures and numerous branches, making it difficult to locate the fault point when a occurs. This article studies precise positioning problem of single-phase grounding faults in rural networks. A new method for locating multi-terminal traveling wave based on principle time information matching is proposed. Firstly, according network structure, database arrival each detection device established advance. Then, after occurs, compared with database,...
Cellular traffic prediction is an indispensable part for intelligent telecommunication networks. Nevertheless, due to the frequent user mobility and complex network scheduling mechanisms, cellular often inherits complicated spatial-temporal patterns, making incredibly challenging. Although recent advanced algorithms such as graph-based approaches have been proposed, they frequently model spatial dependencies based on static or dynamic graphs neglect coexisting multiple correlations induced...
Mobile network traffic forecasting is one of the key functions in daily operation. A commercial mobile large, heterogeneous, complex and dynamic. These intrinsic features make far from being solved even with recent advanced algorithms such as graph convolutional network-based prediction approaches various attention mechanisms, which have been proved successful vehicle forecasting. In this paper, we cast problem a spatial-temporal sequence task. We propose novel deep learning architecture,...
This paper is concerned with TV recommendation, where one major challenge the coupling behavior issue that behaviors of multiple users are coupled together and not directly distinguishable because share same account. Unable to identify current watching user use could lead sub-optimal recommendation results due noise introduced by other users. Most existing methods deal this either unsupervised clustering algorithms or depending on latent representation learning strong assumptions. However,...
Calibration in recommender systems ensures that the user's interests distribution over groups of items is reflected with their corresponding proportions recommendation, which has gained increasing attention recently. For example, a user who watched 80 entertainment videos and 20 knowledge expected to receive recommendations comprising about 80% 20% as well. However, calls for responsible it become inadequate just match users' historical behaviors especially when are grouped by qualities,...
As an efficient tool for approximate similarity computation and search, Locality Sensitive Hashing (LSH) has been widely used in many research areas including databases, data mining, information retrieval, machine learning. Classical LSH methods typically require to perform hundreds or even thousands of hashing operations when computing the sketch each input item (e.g., a set vector); however, this complexity is still too expensive impractical applications requiring processing real-time. To...
Computer-aided diagnosis (CAD) has become a major research topic in medical imaging, and one of the most important CAD applications is detection lung nodules. The paper to develop system for automatically detecting nodules computed tomography (CT) images. includes three parts: pulmonary parenchyma segmentation, ROI extraction, nodule prediction based on ADE-Co-Forest. At beginning, we proposed new segmentation method; In stage circle shape descriptor exploited reduce false positives;...
Urban traffic state analysis plays an important role in the solution of congestion problem. To estimate effectively is a foundational work for improving condition and preventing congestion. In this paper, novel pattern-based approach proposed to model clustering classification state. First, fuzzy-set method utilized divide into number patterns. Then multiclass support vector machine (MSVM) applied these states with real-time data. The result shows that promising dynamic estimation road can...
Accurate traffic flow forecasting is key to the development of intelligent transportation systems (ITS). The support vector regression (SVR) method employed for and comparative results between SVR BP model using real data SCOOT system in Dalian city also presented this paper. Since machines have better generalization performance can guarantee global minima given training data, it believed that will perform well real-time forecasting. However, good highly depends on parameter selection (PS)....
Many applications in real life can produce a large amount of data which be modeled by graph. A graph usually has millions vertices and billions edges. This paper presents BSP-based system, called BC-BSP+, to process graphs iteratively parallel. It the flexibility configure policies (i.e., disk management parameters) extend functions programming interfaces), compute large-scale graphs, tolerate faults, balance loads. Especially, three partition strategies BC-BSP+ are proposed support...