A5103829995- Time Series Analysis and Forecasting
- Network Traffic and Congestion Control
- Machine Learning in Healthcare
- Anomaly Detection Techniques and Applications
- Speech Recognition and Synthesis
- Speech and Audio Processing
- Internet Traffic Analysis and Secure E-voting
- Traffic Prediction and Management Techniques
- Music and Audio Processing
- Network Security and Intrusion Detection
- Artificial Intelligence in Healthcare
- Sepsis Diagnosis and Treatment
- Data Stream Mining Techniques
- Video Coding and Compression Technologies
- Air Quality and Health Impacts
- Energy Load and Power Forecasting
- Complex Systems and Time Series Analysis
- Cardiac, Anesthesia and Surgical Outcomes
- Image and Video Quality Assessment
- Advanced MIMO Systems Optimization
- Millimeter-Wave Propagation and Modeling
- Natural Language Processing Techniques
- Peer-to-Peer Network Technologies
- Allergic Rhinitis and Sensitization
- Software-Defined Networks and 5G
Purple Mountain Laboratories
2019-2023
Handong Global University
2022
Korea Post
2022
Pohang University of Science and Technology
2022
Washington University in St. Louis
2022
Keimyung University
2022
Electronics and Telecommunications Research Institute
2022
Nanjing Institute of Industry Technology
2020-2021
The University of Queensland
2020
University of Science and Technology of China
2016
Renal dysfunction, which is associated with bad clinical outcomes, one of the most common complications heart failure (HF). Timely prediction renal dysfunction can help medical staffs intervene early to avoid catastrophic consequences. In this paper, we proposed a multi-task deep and wide neural network (MT-DWNN) for predicting fatal during hospitalization. The algorithm was tested on dataset collected from Chinese PLA General Hospital, contains 35,101 hospitalizations HF diagnosis last 18...
Time series anomaly detection through unsupervised methods has been an active research area in recent years due to its enormous potential for networks management. The representation and reconstruction of time have made extraordinary progress existing works. However, is known be complex terms their temporal dependency stochasticity, which makes difficult. To this end, we propose a novel approach based on decomposition auto-transformer networks(DATN) detection. decomposed into seasonal trend...
With the rapid development of Internet, various Internet based applications have emerged, and data traffic on has increased dramatically. Network congestion become a core problem in field network research. Appropriate control (CC) can alleviate problems. Due to complexity congestion, is very difficult problem. The existing scheme still cannot effectively solve This article introduces some important algorithms TCP, points out advantages disadvantages these through comparison. Moreover, we...
Multiple organ dysfunction syndrome (MODS) is one of the leading causes death in critically ill patients. MODS result a dysregulated inflammatory response that can be triggered by various causes. Owing to lack an effective treatment for patients with MODS, early identification and intervention are most strategies. Therefore, we have developed variety warning models whose prediction results interpreted Kernel SHapley Additive exPlanations (Kernel-SHAP) reversed diverse counterfactual...
Deep learning (DL)-based speaker diarization methods have proven powerful performance comparing to traditional clustering-based for multi-talker speech and recognition in farfield scenes. However, most DL-based approaches cannot utilize the spatial information well due poor robustness unknown array topology acoustic scenario. In this paper, a long-term iterative mask estimation (SLT-IME) method is proposed improve of various real-world scenarios. First, complex angular central gaussian...
We propose a novel speaker-dependent (SD) approach to joint training of deep neural networks (DNNs) with an explicit speech separation structure for multi-talker recognition in single-channel setting. First, multi-condition strategy is designed SD-DNN recognizer scenarios, which can significantly reduce the decoding runtime and improve accuracy over approaches that use speaker-independent DNN models complicated framework. In addition, SD regression mapping acoustic features mixed target...
Cloud radio access networks (CRAN) has become an excellent network architecture in the fifth-generation wireless communication systems, which can highly increase capacity and coverage compared with conventional networks. However, order to provide best services, appropriate resource management must be applied. This article considers transmission scheduling of downlink OFDMA-based Millimeter-Wave (mmWave) CRAN, where data on each subcarrier (SC) used by different remote heads (RRHs). A joint...
Cohort study is one of the most commonly used methods in medical and public health researches, which result longitudinal data. Conventional statistical models machine learning are not capable modeling evolution trend variables In this article, we propose a Trend Analysis Neural Networks (TANN), by adaptive feature learning. TANN was tested on dataset Kaiuan research. The task to predict occurrence cardiovascular events within 2 5 years, with three repeated examinations during 2008 2013. For...
Abstract Background Predicting patient mortality risk facilitates early intervention in intensive care unit (ICU) patients at greater of disease progression. This study applies machine learning methods to multidimensional clinical data dynamically predict ICU patients. Methods A total 33,798 the MIMIC-III database were collected. An integrated model NIMRF (Network Integrating Memory Module and Random Forest) based on variables such as vital sign laboratory was developed death for four non...
Accurate and timely prediction of the Internet traffic flow is important for network performance improvement. Few methods incorporate topology information networks. In this paper, we present a novel forecasting algorithm named TTGCN, which applies graph neural networks on each link backbone network. The was represented by adjacency matrix, models relationship between links. design makes TTGCN capable capturing both temporal topological flow. validated UKERNA, dataset captured from real...
Unsupervised anomaly detection for time series has been an active research area due to its enormous potential wireless network management. Existing works have made extraordinary progress in representation, reconstruction and forecasting. However, long-term temporal patterns prohibit the model from learning reliable dependencies. To this end, we propose a novel approach based on Transformer with dilated convolution detection. Specifically, provide module extract dependence features. Extensive...
Cohort study is one of the most commonly used methods in medical and public health researches, which result longitudinal data. Conventional statistical models machine learning are not capable modeling evolution trend variables In this paper, we propose a Trend Analysis Neural Networks (TANN), by adaptive feature learning. TANN was tested on dataset Kaiuan research. The task to predict occurrence death within 5 years, with 3 repeated examinations from 2008 2013. AUC 0.7888, slightly...
This technical report details our submission system to the CHiME-7 DASR Challenge, which focuses on speaker diarization and speech recognition under complex multi-speaker scenarios. Additionally, it also evaluates efficiency of systems in handling diverse array devices. To address these issues, we implemented an end-to-end introduced a rectification strategy based multi-channel spatial information. approach significantly diminished word error rates (WER). In terms recognition, utilized...
The unsupervised anomaly detection in KPI (Key Performance Indicator) series has been an active research area due to its enormous potential for application industry. representation, reconstruction, and forecasting have made extraordinary progress existing work. However, long-term temporal patterns prohibit the model from learning reliable dependencies. To this end, we propose a novel approach based on VAE-TCN hybrid model. Our uses VAE (variational automatic coder) learn robust local...