A5100769232- Speech Recognition and Synthesis
- Alzheimer's disease research and treatments
- Neuroinflammation and Neurodegeneration Mechanisms
- Speech and Audio Processing
- Endometriosis Research and Treatment
- Software Engineering Research
- Advanced Image Processing Techniques
- Software Testing and Debugging Techniques
- Advanced Neuroimaging Techniques and Applications
- EEG and Brain-Computer Interfaces
- Uterine Myomas and Treatments
- Advanced Vision and Imaging
- Machine Learning in Healthcare
- Topic Modeling
- Solar Radiation and Photovoltaics
- Music and Audio Processing
- Gynecological conditions and treatments
- Energy Load and Power Forecasting
- Advanced Graph Neural Networks
- Face and Expression Recognition
- Software System Performance and Reliability
- Natural Language Processing Techniques
- Optical Network Technologies
- Thermoelastic and Magnetoelastic Phenomena
- 3D Surveying and Cultural Heritage
Washington University in St. Louis
2021-2024
Peng Cheng Laboratory
2022-2024
Beijing University of Posts and Telecommunications
2023-2024
Mallinckrodt (United States)
2022-2024
Nanjing University of Information Science and Technology
2024
Chinese Academy of Sciences
2022-2023
Xi'an Technological University
2022-2023
Zhejiang University
2023
Shenzhen Institutes of Advanced Technology
2022-2023
Central China Normal University
2023
Guangtao Zeng, Wenmian Yang, Zeqian Ju, Yue Sicheng Wang, Ruisi Zhang, Meng Zhou, Jiaqi Xiangyu Dong, Ruoyu Hongchao Fang, Penghui Zhu, Shu Chen, Pengtao Xie. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). 2020.
Alzheimer disease (AD) is primarily associated with accumulations of amyloid plaques and tau tangles in gray matter, however, it now acknowledged that neuroinflammation, particularly white matter (WM), significantly contributes to the development progression AD. This study aims investigate WM neuroinflammation continuum AD its association pathologies cognition using diffusion-based imaging (NII).
This study aims to quantify microglial activation in individuals with Alzheimer disease (AD) using the 18-kDa translocator protein (TSPO) PET imaging hippocampus and precuneus, 2 AD-vulnerable regions, evaluate association of baseline neuroinflammation amyloidosis, tau, longitudinal cognitive decline.Twenty-four participants from Knight Disease Research Center (Knight ADRC) were enrolled classified into stable cognitively normal, progressor, symptomatic AD groups based on clinical dementia...
Abstract Uterine peristalsis, characterized by spontaneous slow-wave contractions of the subendometrial layer uterine myometrium, occurs throughout menstrual cycle. Disruptions in peristalsis patterns may occur women experiencing abnormal bleeding, endometriosis, and infertility. Current tools to measure humans have limitations that hamper their research or clinical utility. Here, we developed an electrophysiological imaging system with wearable electrical sensors noninvasively quantify...
Abstract Electromyometrial imaging (EMMI) was recently developed to image the three-dimensional (3D) uterine electrical activation during contractions noninvasively and accurately in sheep. Herein we describe development application of a human EMMI system evaluate 3D patterns at high spatial temporal resolution term labor. We demonstrate successful integration subjects’ clinical visits generate surface potential maps, electrograms, sequence through an inverse solution using up 192 electrodes...
<title>Abstract</title> Uterine peristalsis involves spontaneous, mild contractions of the sub-endometrial layer in nonpregnant uterus, playing a crucial role menstruation and fertility. Disruptions peristaltic patterns have been linked to conditions like subfertility, abnormal uterine bleeding, endometriosis. Peristalsis Imaging (UPI) was developed noninvasively image 3D electrical activation with high spatial temporal resolution. However, UPI cannot capture mechanical deformation or...
Self-homodyne mode division multiplexing (SH-MDM) is considered a promising candidate for achieving high capacity and low complexity in next-generation optical interconnects. However, practical implementation remains challenging with channel impairments caused by transmission mechanisms, particularly coupling (MC) differential group delay (DMGD), as well receiver hardware (RHI) originating from imperfect coherent receivers, including IQ skew, amplitude imbalance, phase imbalance. To address...
Lung vessel segmentation of computed tomography (CT) images is important in clinical practise and challenging due to difficulties associated with minor size blurred edges lung vessels. A method proposed for based on a random forest classifier sparse auto‐encoder features. First, the multi‐scale representations are obtained using Gaussian pyramid. Second, three layers trained randomly selected patches these images. Next, weight used as convolution kernel extract features different scale...
Diabetic cardiomyopathy is one of the most life-threatening diabetic complications. However, previous studies only discuss a particular aspect or characteristic DCM, current state and trends were explored by limited research. We aimed to perform systemically bibliometric study DCM research progress status in past decade, visualize internal conceptual structure potential associations, further explore prospective trends.Articles related published from January 2012 December 2021 collected Web...
In this paper, we propose a cross-modal student-teacher learning framework to make full use of externally abundant acoustic data in addition given task-specific audio-visual training database for improving speech recognition performance under the low signal-to-noise-ratio (SNR) and mismatch conditions. First, teacher model is trained with large-sized audio-only databases. Next, student, namely deep neural network (DNN) model, on small-sized minimize Kullback-Leibler (KL) divergence between...
We propose an environment adaptation approach that improves deep speech enhancement models via minimizing the Kullback-Leibler divergence between posterior probabilities produced by a multi-condition senone classifier (teacher) fed with noisy features and clean-condition (student) enhanced to transfer existing neural network (DNN) enhancer specific environments without using noisy/clean paired target waveforms needed in conventional DNN-based spectral regression. Our solution not only...
Spatial structured light (SL) can achieve three-dimensional measurements with a single shot. As an important branch in the field of dynamic reconstruction, its accuracy, robustness, and density are vital importance. Currently, there is wide performance gap spatial SL between dense reconstruction (but less accurate, e.g., speckle-based SL) accurate often sparser, shape-coded SL). The central problem lies coding strategy designed features. This paper aims to improve quantity reconstructed...
Deep learning frameworks (DLFs) have been playing an increasingly important role in this intelligence age since they act as a basic infrastructure for wide range of AI-based applications. Meanwhile, multi-programming-language (MPL) software systems, DLFs are inevitably suffering from bugs caused by the use multiple programming languages (PLs). Hence, it is paramount significance to understand (especially involving PLs, i.e., MPL bugs) DLFs, which can provide foundation preventing, detecting,...
Leveraging upon transfer learning, we distill the knowledge in a conventional wide and deep neural network (DNN) into narrower yet deeper model with fewer parameters comparable system performance for speech enhancement. We present three transfer-learning solutions to accomplish our goal. First, embedded form of output values high-performance DNN is used guide training smaller sequential learning. In second multi-task learning solution, trained learn value larger DNN, enhancement task...
Several recent works discussed application-driven image restoration neural networks, which are capable of not only removing noise in images but also preserving their semantic-aware details, making them suitable for various high-level computer vision tasks as the pre-processing step. However, such approaches require extra annotations tasks, order to train joint pipeline using hybrid losses. The availability those is yet often limited a few sets, potentially restricting general applicability...
Throughout the menstrual cycle, spontaneous mild contractions in inner layer of uterine smooth muscle cause peristalsis, which plays a critical role normal menstruation and fertility. Disruptions peristalsis patterns may occur women experiencing subfertility, abnormal bleeding, ovulatory dysfunction, endometriosis, other disorders. However, current tools to measure humans have limitations that hamper their research or clinical utilities. Here, we describe an electrophysiological imaging...
Pancreatic ductal adenocarcinoma (PDAC) is not sensitive to immune checkpoint blockade therapy, and negative feedback of tumor evasion might be partly responsible. We isolated CD8+ T cells cultured them in vitro. Proteomics analysis was performed compare changes Panc02 cell lines with conditioned medium, leucine-rich repeat kinase 2 (LRRK2) identified as a differential gene. LRRK2 expression related spatial distribution PDAC clinical samples upregulated by via interferon gamma (IFN-γ)...
We propose mixed-precision integer-arithmetic-only NNs for low-complexity equalization in a 50-Gb/s 25-km DML-based PAM4 IM/DD link based on input neuron partitioning. The proposed scheme avoids floating-point weight operations and saves 26.3% memory at minimum NN size compared with fixed-precision quantization counterpart without degrading BER.
Medical dialogue systems are promising in assisting telemedicine to increase access healthcare services, improve the quality of patient care, and reduce medical costs. To facilitate research development systems, we build two large-scale datasets: MedDialog-EN MedDialog-CN. is an English dataset containing 0.3 million conversations between patients doctors 0.5 utterances. MedDialog-CN Chinese 1.1 4 our best knowledge, MedDialog-(EN,CN) largest datasets date. The available at...