Search

Photoplethysmography (PPG) provides a low-cost, non-invasive method to continuously monitor various cardiovascular parameters. PPG signals are generated by wearable devices and frequently contain large artifacts caused by external factors, such as motion of the human subject. In order to ensure robust and accurate extraction of physiological parameters, corrupted areas of the signal need to be identified and handled appropriately. Previous methodology relied either on handcrafted feature detectors or signal metrics which yield sub-optimal performance, or relied on machine learning techniques such as deep neural networks (DNN) which lack interpretability and are computationally and memory intensive. In this work, we present a novel method to learn a small set of interpretable convolutional kernels that has performance similar to -- and often better than -- the state-of-the-art DNN approach with several orders of magnitude fewer parameters. This work allows for efficient, robust, and interpretable signal quality assessment and artifact segmentation on low-power devices., 16 pages, 6 figures

Retrieval-augmented methods have received increasing attention to support downstream tasks by leveraging useful information from external resources. Recent studies mainly focus on exploring retrieval to solve knowledge-intensive (KI) tasks. However, the potential of retrieval for most non-knowledge-intensive (NKI) tasks remains under-explored. There are two main challenges to leveraging retrieval-augmented methods for NKI tasks: 1) the demand for diverse relevance score functions and 2) the dilemma between training cost and task performance. To address these challenges, we propose a two-stage framework for NKI tasks, named PGRA. In the first stage, we adopt a task-agnostic retriever to build a shared static index and select candidate evidence efficiently. In the second stage, we design a prompt-guided reranker to rerank the nearest evidence according to task-specific relevance for the reader. Experimental results show that PGRA outperforms other state-of-the-art retrieval-augmented methods. Our analyses further investigate the influence factors to model performance and demonstrate the generality of PGRA. Codes are available at https://github.com/THUNLP-MT/PGRA.

In intensive care units (ICUs), critically ill patients are monitored with electroencephalograms (EEGs) to prevent serious brain injury. The number of patients who can be monitored is constrained by the availability of trained physicians to read EEGs, and EEG interpretation can be subjective and prone to inter-observer variability. Automated deep learning systems for EEG could reduce human bias and accelerate the diagnostic process. However, black box deep learning models are untrustworthy, difficult to troubleshoot, and lack accountability in real-world applications, leading to a lack of trust and adoption by clinicians. To address these challenges, we propose a novel interpretable deep learning model that not only predicts the presence of harmful brainwave patterns but also provides high-quality case-based explanations of its decisions. Our model performs better than the corresponding black box model, despite being constrained to be interpretable. The learned 2D embedded space provides the first global overview of the structure of ictal-interictal-injury continuum brainwave patterns. The ability to understand how our model arrived at its decisions will not only help clinicians to diagnose and treat harmful brain activities more accurately but also increase their trust and adoption of machine learning models in clinical practice; this could be an integral component of the ICU neurologists' standard workflow., 20 pages including appendices, 7 figures, submitted for peer review

Obtaining large-scale well-annotated is always a daunting challenge, especially in the medical research domain because of the shortage of domain expert. Instead of human annotation, in this work, we use the alarm information generated from bed-side monitor to get the pseudo label for the co-current photoplethysmography (PPG) signal. Based on this strategy, we end up with over 8 million 30-second PPG segment. To solve the label noise caused by false alarms, we propose the cluster consistency, which use an unsupervised auto-encoder (hence not subject to label noise) approach to cluster training samples into a finite number of clusters. Then the learned cluster membership is used in the subsequent supervised learning phase to force the distance in the latent space of samples in the same cluster to be small while that of samples in different clusters to be big. In the experiment, we compare with the state-of-the-art algorithms and test on external datasets. The results show the superiority of our method in both classification performance and efficiency.

A graphene aerogel-coated molybdenum trioxide/polypyrrole (MoO3/GA@PPy) composite was synthesized via hydrothermal and in site polymerization reactions. The character of PPy, MoO3/GA, and MoO3/GA@PPy were investigated by X-ray diffraction, Fourier transform spectroscopy, Raman spectroscopy, photoluminescence spectroscopy, scanning electron microscopy, and X-ray photoelectron spectroscopy. The results of the morphological and structural analysis clearly revealed the formation of ternary nanocomposites. The electrochemical properties, such as capacitance, specific capacitance retention at different current loads, cycling stability, and impedance, were discussed as well. Improvement of electrochemical performances of the ternary nanocomposites was observed. The MoO3/GA@PPy showed a maximum specific capacitance of 1788 F/g at current density of 1 A/g.

We propose a balanced coarsening scheme for multilevel hypergraph partitioning. In addition, an initial partitioning algorithm is designed to improve the quality of k-way hypergraph partitioning. By assigning vertex weights through the LPT algorithm, we generate a prior hypergraph under a relaxed balance constraint. With the prior hypergraph, we have defined the Wasserstein discrepancy to coordinate the optimal transport of coarsening process. And the optimal transport matrix is solved by Sinkhorn algorithm. Our coarsening scheme fully takes into account the minimization of connectivity metric (objective function). For the initial partitioning stage, we define a normalized cut function induced by Fiedler vector, which is theoretically proved to be a concave function. Thereby, a three-point algorithm is designed to find the best cut under the balance constraint., There are some errors in the paper

Supplement 1

The dynamic process of polymorphic transformations provides deep insights into the understanding of both the performance of explosives and its packing sensitivity. While terahertz spectroscopy is a sensitive tool that can be used to characterize its intermolecular forces. In this paper, terahertz time-domain spectroscopy (THz-TDS) is employed to investigate the polymorphic transformations of the explosive 2,2dinitroethylene-1,1-diamine (FOX-7) under a heating process. Solid-state density functional theory (DFT) uncovers the physical mechanisms that contribute to the polymorphic transformation. This is achieved by assessing the calculated spectra relative to the actual spectra from experiment. The results suggest that the THz spectral features that arise represent a redistribution of the intermolecular forces during the polymorphic transformations. This research will assist the understanding of the evolving properties of FOX-7, under external stimuli, which is vital knowledge for an explosive substance that can combust or detonate.

The principle, characteristics, operation mode and performance test results of the evaporative cooling air–water air-conditioning system are introduced in this paper, in which the evaporative water chiller is adopted as the main cooling source and the evaporative cooling fresh air handling unit as the auxiliary cooling source in a data center room in Xinjiang. The system utilizes local dry air energy and low-temperature air as natural cooling sources. It not only realizes 100% free cooling throughout the year, but also effectively solves the high-temperature crash problem and freezing in the cooling tower of traditional water-cooled air-conditioning system in winter in Xinjiang based on direct evaporative cooling technology, indirect evaporative cooling technology and ethylene glycol-free cooling technology. According to the summer test of the system, the system is stable, and the refrigerating coefficient of performance of the air-conditioning system is 6.65, and the summated refrigerating coefficient of performance (SCOP) is 16.64, which has a significant effect on reducing the energy consumption of data center.

In order to improve the accuracy of foreground target detection and establish a stable background model, this paper proposes a multi-feature fusion background modeling algorithm, which initializes the background model with the spatial correlation between the first frame pixel and the domain pixel, and quickly establishes the background. model. A multi-feature sample set consisting of pixel values, update frequency, update time, and adaptive dynamic coefficients is updated with temporal correlation of subsequent intra-pixels. According to the multi-feature sample set, the background complexity is adjusted to adjust the update speed of the model in different regions, which effectively improves the ghost phenomenon of the foreground target and reduces the false holes in the target and the false foreground in the background. The test results of multiple sets of data sets show that the proposed algorithm improves the adaptability and robustness of foreground target detection in scenarios with high dynamic changes.

Haiyuan Yu,1,* Jun Pan,1,* Zhicheng Guo,1 Chunhua Yang,1,2 Lijun Mao1,2 1Department of Urinary Surgery, the Affiliated Hospital of Xuzhou Medical University, Xuzhou 221000, China; 2Jiangsu Key Laboratory of Biological Cancer Therapy, Department of Clinic Institute, Xuzhou Medical University, Xuzhou 221002, China *These authors contributed equally to this work Abstract: In recent years, the, chimeric antigen receptor T (CAR-T) cell therapy as an adoptive immunotherapy has received great attention and made great breakthroughs. CAR-T cells show great specificity, targeting, and less major histocompatibility complex restriction in tumor immunotherapy, significantly different from traditional T cells. In spite of the progress of CART-T technology in the treatment of lymphoma, leukemia, and other blood system tumor, there are still many difficulties in the treatment of solid tumors by CAR-T technology. In this review, we will make a brief summary of the present situation of CAR-T cells in the treatment of prostate cancer, and discuss the promise of the application of this technology to prostate cancer therapy. Keywords: prostate cancer, CAR-T cells, adoptive immunotherapy, review