Search

Background: Rhabdomyolysis (RM) associated acute kidney injury (AKI) is the most common systemic complication of RM. The present study aimed to assess the clinical characteristics and risk factors for intensive care unit (ICU) transfer for patients with RM-associated AKI.Methods: We included all patients who were age ≥ 18 years old with a diagnosis of RM from September 2012 to October 2018 and divided them into RM-associated AKI group and RM without AKI group. The primary outcome was transferring to ICU treatment. Regression analysis was performed to identify factors associated with ICU treatment and recovery of renal function.Results: Among the 149 patients with RM, 68 (45.6%) developed AKI. The percentage of patients with AKI who transferred to ICU was higher than patients without AKI (33.8% versus 12.3%, P < 0.002). Additionally, patients with AKI had higher percentage of undergoing dialysis (19.1% versus 2.5%, P < 0.01), all-cause mortality (13.2% versus 1.2%, P < 0.01), cost of hospitalization [10.8 1,000 yuan, IQR (5.5, 3.5) versus 5.9 1,000 yuan, IQR 5.9 (3.6, 9.9), P = 0.03], as well as longer length of hospital stay [8.0 (5.0, 14.0)] versus [6.0 (4.0, 11.0)], P = 0.02). Moreover, most patients with AKI achieved complete recovery (77.9%) at discharge. After adjusting for potential risk factors, RM-associated AKI remained an independent risk factor for ICU transfer (OR = 3.0, 95% CI, 1.11–8.3, P = 0.03). However, ICU transfer was not associated with recovery of renal function (OR = 0.88, 95% CI, 0.22–3.57, P = 0.856).Conclusion: RM leaded to AKI in most patients. RM-associated AKI could cause worse clinical outcome and predict ICU transfer for patients with RM.

Graph convolutional network (GCN) emerges as a promising direction to learn the inductive representation in graph data commonly used in widespread applications, such as E-commerce, social networks, and knowledge graphs. However, learning from graphs is non-trivial because of its mixed computation model involving both graph analytics and neural network computing. To this end, we decompose the GCN learning into two hierarchical paradigms: graph-level and node-level computing. Such a hierarchical paradigm facilitates the software and hardware accelerations for GCN learning. We propose a lightweight graph reordering methodology, incorporated with a GCN accelerator architecture that equips a customized cache design to fully utilize the graph-level data reuse. We also propose a mapping methodology aware of data reuse and task-level parallelism to handle various graphs inputs effectively. Results show that Rubik accelerator design improves energy efficiency by 26.3x to 1375.2x than GPU platforms across different datasets and GCN models.

In recent years, tensor computation has become a promising tool for solving big data analysis, machine learning, medical image and EDA problems. To ease the memory and computation intensity of tensor processing, decomposition techniques, especially Tensor-train Decomposition(TTD), are widely adopted to compress the extremely high-dimensional tensor data. Despite TTD’s potential to break the curse of dimensionality, researchers have not yet leveraged its full computational potential, mainly because of two reasons:(1) Executing TTD itself is time-and energy-consuming due to the singular value decomposition(SVD) operation inside each of TTD’s iteration; (2) Additional software/hardware optimizations are often required to process the obtained TT-format data in certain applications such as deep learning inference. In this paper, we address these challenges with two approaches. Firstly, we propose an algorithm-hardware co-design with customized architecture namely TTD Engine to accelerate TTD. We use MRI image compression as a demo application to illustrate the efficacy of the proposed accelerator. Secondly, we present a case study demonstrating the benefit of TT-format data processing and the efficacy of using TTD Engine. In the case study, we use TT approach to realize convolution operation, which is difficult and nontrivial for TT-format data. Experimental results show that, TTD Engine achieves, on average, 14.9×∼36.9× speedup over CPU implementations and 4.1×∼9.9× speedup compared to the GPU baseline. The energy efficiency is also improved by at least 14.4× and 5.4× over CPU and GPU, respectively. Moreover, our hardware-enabled TT-format data processing further leads to more efficient implementations of complicated operations and applications.

The rapid progress in the development of non-fullerene electron acceptors has led to key breakthroughs in the power conversion efficiency of organic solar cells (OSCs). Nevertheless, the development of non-fullerene OSCs are still hindered by the short exciton diffusion length in both donor and acceptor domains, which is paramount in determining the charge generation properties in OSCs. Reducing the domain size to match the diffusion length via bulk heterojunction (BHJ) structure has been an essential strategy to overcome this limitation, but will inevitably compromise charge collection due to the intermixing of donor and acceptor. Here, we use a solution-based contact film transfer method to fabricate planar heterojunction (PHJ) device, which serves as an excellent platform to investigate exciton diffusion and charge collection owing to the clearly defined interface. By reconstructing external quantum efficiency spectra and device modeling in PHJ cells based on a large bandgap polymer J52 and a low bandgap small molecule IEICO-4F, exciton diffusion lengths of 9.2 nm and 7.5 nm are extracted for J52 and IEICO-4F, respectively. The experimentally obtained values for exciton diffusion length confirm that BHJ structure is an important prerequisite for efficient organic solar cells. Nevertheless, as revealed by the carrier dynamic studies, BHJ devices could suffer from severe recombination losses as a result of energetic disorder and traps. By considering the recombination of free and trapped carriers and intermediates in the photocurrent generation processes, we show that the stronger recombination of BHJ compared to PHJ is also related to its shorter CT states lifetime and lower CT states density based on the reduced Langevin recombination analysis. The inherent limitation in BHJ devices is thus a major disadvantage that need to be overcome for further improvement. On the basis of the results of exciton diffusion length measurement and carrier dynamic study, we propose that finely tuning domain size in non-fullerene OSCs still play a crucial role in controlling the device performance.

In recent years, plenty of metrics have been proposed to identify networks that are free of gradient explosion and vanishing. However, due to the diversity of network components and complex serial-parallel hybrid connections in modern DNNs, the evaluation of existing metrics usually requires strong assumptions, complex statistical analysis, or has limited application fields, which constraints their spread in the community. In this paper, inspired by the Gradient Norm Equality and dynamical isometry, we first propose a novel metric called Block Dynamical Isometry, which measures the change of gradient norm in individual block. Because our Block Dynamical Isometry is norm-based, its evaluation needs weaker assumptions compared with the original dynamical isometry. To mitigate the challenging derivation, we propose a highly modularized statistical framework based on free probability. Our framework includes several key theorems to handle complex serial-parallel hybrid connections and a library to cover the diversity of network components. Besides, several sufficient prerequisites are provided. Powered by our metric and framework, we analyze extensive initialization, normalization, and network structures. We find that Gradient Norm Equality is a universal philosophy behind them. Then, we improve some existing methods based on our analysis, including an activation function selection strategy for initialization techniques, a new configuration for weight normalization, and a depth-aware way to derive coefficients in SeLU. Moreover, we propose a novel normalization technique named second moment normalization, which is theoretically 30% faster than batch normalization without accuracy loss. Last but not least, our conclusions and methods are evidenced by extensive experiments on multiple models over CIFAR10 and ImageNet., Comment: Under review as a regular paper in TPAMI

The rapid economic development and climate change have accelerated the changes in China’s food production and have a potential impact on food security. In this paper, the grain sown area from 2001 to 2019 was selected to analyze the spatio-temporal evolution and driving factors of China’s grain production through spatial autocorrelation analysis and geographically weighted regression. Our findings were as follows: (1) From the perspective of time characteristics, China’s grain production from 2001 to 2019 experienced four stages, rapid decline, rapid growth, steady growth, and slow decline, although with an overall upward trend. (2) From the perspective of spatial characteristics, the overall spatial pattern had a significant positive correlation. The high values were mainly concentrated in Shandong, Anhui, and Jilin and moved to the northeast China as time went on. (3) In terms of influencing factors, the positive impact of agricultural labor force on the grain production gradually decreased, showing a decreasing trend from southwest to northeast. The promotion of agricultural mechanization on the grain production increased year by year, with the spatial distribution characteristics of high in the northeast and low in the southwest. Besides, the coefficient of water resources endowment was negative, showing a spatial distribution pattern of high north and low south.

In this paper, a novel D eep M ulti-view J oint C lustering ( DMJC ) framework is proposed, where multiple deep embedded features, multi-view fusion mechanism, and clustering assignments can be learned simultaneously. Through the joint learning strategy, the clustering-friendly multi-view features and useful multi-view complementary information can be exploited effectively to improve the clustering performance. Under the proposed joint learning framework, we design two ingenious variants of deep multi-view joint clustering models, whose multi-view fusion is implemented by two kinds of simple yet effective schemes. The first model, called DMJC-S, performs multi-view fusion in an implicit way via a novel multi-view soft assignment distribution. The second model, termed DMJC-T, defines a novel multi-view auxiliary target distribution to conduct the multi-view fusion explicitly. Both DMJC-S and DMJC-T are optimized under a KL divergence objective. Experiments on eight challenging image datasets demonstrate the superiority of both DMJC-S and DMJC-T over single/multi-view baselines and the state-of-the-art multi-view clustering methods, which proves the effectiveness of the proposed DMJC framework. To the best of our knowledge, this is the first work to model the multi-view clustering in a deep joint framework, which will provide a meaningful thinking in unsupervised multi-view learning.

Hundreds of C60 derivatives stand out as electrontransporting materials(ETMs), for example, in perovskite solar cells(PSCs), due to their properties on electron extraction or defect passivation. However, it still lacks of guidelines to update C60-based ETMs with excellent photoelectric properties. In this work, crystallographic data of eight C60-based ETMs, including pristine C60 and the well-known PCBM as well as six newly synthesized fullerenes, are analyzed to establish the connections between derivatized structures and photoelectric properties for the typical carbon cluster of C60. In terms of packing centroid-centroid distance between neighboring carbon cages, the crystallographic data are useful for probing photoelectric properties, such as electrochemical properties, electron mobility and photovoltaic performances, and therefore facilitate to design novel C60-based ETMs for PSCs with high performances.

Background Sugarcane is the most important sugar crops in the world. Like other crops, sugarcane is also suffered from herbivorous insect attack. The oriental armyworm Mythimna separata is a devastating pest of various crop in northeast Asia, outbreak of this pest resulted in substantially yield loss including sugarcane. However, the plant defense response situation is widely acquisition in model crops, but little information about how sugarcane plants defend themselves against this herbivore at the molecular and biochemical levels is studied. Results We combined transcriptome and metabolomics analysis to investigate the changes in gene expression and metabolic processes that occurred in sugarcane plants after continuous feeding by M. separata larvae for 12, and 24 h. We identified 13 662 genes and 55 metabolites that were differentially regulated in sugarcane plants fed upon by M. separata. The genes involved in phytohormones, transcription factors, and kinase-related were activated and the metabolism compounds such as carbohydrate, amino acid, ferulic substances and glutathione were detected regulated in sugarcane defense response. Comparable analyses showed a close correspondence relationship among pathways of phenylalanine metabolism, phenylpropanoid biosynthesis, and flavonoid biosynthesis in transcript and metabolite profiles. Furthermore, a bioassay experiment was conducted to test the influence of up-regulated metabolites on M. separata growth and found chlorogenic acid showed lethal effect. Conclusion The results of our study greatly enhanced understanding on sugarcane induced defense response mechanism against herbivore infestation at the transcriptional and metabolic levels. Also make contributions to provide clues for development of green pest control method. This article is protected by copyright. All rights reserved.

We explored the protective effect of spironolactone on cardiac function in the patients undergoing coronary artery bypass grafting (CABG) by determining serum hypoxia-inducible factor-1α (HIF-1α) before and after CABG. We used the propensity score matching method retrospectively to select 174 patients undergoing CABG in our hospital from March 2018 to December 2019. Of the 174 patients, 87 patients taking spironolactone for more than 3 months before CABG were used as a test group and other 87 patients who were not taking spironolactone as a control group. In all patients, serum HIF-1α and troponin I levels were determined before as well as 24 hours and 7 days after CABG, serum N-terminal probrain natriuretic peptide (NT-proBNP) level was determined before as well as 12, 24, and 36 hours after CABG, and electrocardiographic monitoring was performed within 36 hours after CABG. The results indicated that there were no significant differences in the HIF-1α level between the test group and the control group before and 7 days after CABG, but the HIF-1α level was significantly lower in the test group than that in the control group 24 hours after CABG (P < 0.01). The 2 groups were not significantly different in the troponin I level at any time point. There was no significant difference in the serum NT-proBNP level between the test group and the control group before CABG, but NT-proBNP (BNP) levels were all significantly lower in the test group than those in the control group at postoperative 12, 24, and 36 hour time points (all P

Deep neural network (DNN) accelerators are widely deployed in computer vision, speech recognition, and machine translation applications, in which attacks on DNNs have become a growing concern. This article focuses on exploring the implications of hardware Trojan attacks on DNNs. Trojans are one of the most challenging threat models in hardware security where adversaries insert malicious modifications to the original integrated circuits (ICs), leading to malfunction once being triggered. Such attacks can be conducted by adversaries because modern ICs commonly include third-party intellectual property (IP) blocks. Previous studies design hardware Trojans to attack DNNs with the assumption that adversaries have full knowledge or manipulation of the DNN systems’ victim model and toolchain in addition to the hardware platforms, yet such a threat model is strict, limiting their practical adoption. In this article, we propose a memory Trojan methodology that implants the malicious logics merely into the memory controllers of DNN systems without the necessity of toolchain manipulation or accessing to the victim model and thus is feasible for practical uses. Specifically, we locate the input image data among the massive volume of memory traffics based on memory access patterns and propose a Trojan trigger mechanism based on detecting the geometric feature in input images. Extensive experiments show that the proposed trigger mechanism is effective even in the presence of environmental noises and preprocessing operations. Furthermore, we design and implement the payload and verify that the proposed Trojan technique can effectively conduct both untargeted and targeted attacks on DNNs.

Superhydrophobic surfaces have attracted attention due to their hydrophobic, self-cleaning, anti-icing, and oil/water separation properties. The present study used a nonwoven fabric as a flexible substrate and constructed a high-adhesive superhydrophobic surface by coating the fabric with polydimethylsiloxane (PDMS)-co -polymethyl hydrogen siloxane (PMHS) polymers and embedding mesoporous SiO2 nanoparticles. The water contact angle (WCA) of the superhydrophobic surface was up to 165.2° at a PDMS to PMHS ratio of 4:1. The adhesion to deionized water was 99.7 μN. High hydrophobicity was maintained, even after sandpaper abrasions and flowing water impact. The surface was resistant to acid, alkali, brine, strong oxidation, and heavy metal solutions. The coating exhibited anti-icing and oil/water separation properties. This study provides a facile and effective method for constructing multifunctional superhydrophobic coatings on flexible substrates.

We propose P-RT, the first robotic vision runtime framework to efficiently manage dynamic task executions on mobile systems with multiple accelerators as well as on the cloud to achieve better performance and energy savings. With P-RT, we enable a robot to simultaneously perform autonomous navigation.

BACKGROUND AND PURPOSE: In acute stroke patients with large vessel occlusions, it would be helpful to be able to predict the difference in the size and location of the final infarct based on the outcome of reperfusion therapy. Our aim was to demonstrate the value of deep learning–based tissue at risk and ischemic core estimation. We trained deep learning models using a baseline MR image in 3 multicenter trials. MATERIALS AND METHODS: Patients with acute ischemic stroke from 3 multicenter trials were identified and grouped into minimal (≤20%), partial (20%-80%), and major (≥80%) reperfusion status based on 4- to 24-hour follow-up MR imaging if available or into unknown status if not. Attention-gated convolutional neural networks were trained with admission imaging as input and the final infarct as ground truth. We explored 3 approaches: 1) separate: train 2 independent models with patients with minimal and major reperfusion; 2) pretraining: develop a single model using patients with partial and unknown reperfusion, then fine-tune it to create 2 separate models for minimal and major reperfusion; and 3) thresholding: use the current clinical method relying on apparent diffusion coefficient and time-to-maximum of the residue function maps. Models were evaluated using area under the curve, the Dice score coefficient, and lesion volume difference. RESULTS: Two hundred thirty-seven patients were included (minimal, major, partial, and unknown reperfusion: n = 52, 80, 57, and 48, respectively). The pretraining approach achieved the highest median Dice score coefficient (tissue at risk = 0.60, interquartile range, 0.43–0.70; core = 0.57, interquartile range, 0.30–0.69). This was higher than the separate approach (tissue at risk = 0.55; interquartile range, 0.41–0.69; P = .01; core = 0.49; interquartile range, 0.35–0.66; P = .04) or thresholding (tissue at risk = 0.56; interquartile range, 0.42–0.65; P = .008; core = 0.46; interquartile range, 0.16–0.54; P < .001). CONCLUSIONS: Deep learning models with fine-tuning lead to better performance for predicting tissue at risk and ischemic core, outperforming conventional thresholding methods.

When deploying deep neural networks (DNNs) onto deep learning processors, we usually exploit mixed-precision quantization and voltage–frequency scaling to make tradeoffs among accuracy, latency, and energy. Conventional methods usually determine the quantization–voltage–frequency (QVF) policy before DNNs are deployed onto local devices. However, they are difficult to make optimal customizations for local user scenarios. In this article, we solve the problem by enabling on-device QVF tuning with a new deep learning processor architecture Evolver. Evolver has a QVF tuning mode to deploy DNNs with local customizations before normal execution. In this mode, Evolver uses reinforcement learning to search the optimal QVF policy based on direct hardware feedbacks from the chip itself. After that, Evolver runs the newly quantized DNN inference under the searched voltage and frequency. To improve the performance and energy efficiency of both training and inference, we introduce bidirectional speculation and runtime reconfiguration techniques into the architecture. To the best of our knowledge, Evolver is the first deep learning processor that utilizes on-device QVF tuning to achieve both customized and optimal DNN deployment.

The screening and treatment of laryngeal squamous cell carcinoma (LSCC) still perplexes clinicians, making it necessary to explore new markers. To this end, this research examined the underlying molecular mechanism of LSCC based on high-throughput datasets (n = 249) from multiple databases. It also identified transcription factors (TFs) independently associated with LSCC prognosis. Through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses, differential expression genes of LSCC were deemed relevant to the extracellular matrix and its related structures or pathways, suggesting that the extracellular matrix plays an important role in LSCC. At the same time, several hub genes that may also have important roles in LSCC were identified via protein–protein interaction analysis, including CDC45, TPX2, AURKA, KIF2C, NUF, MUC1, MUC7, MUC4, MUC15, and MUC21. Eight unreported LSCC prognostic TFs – BCAT1, CHD4, FOXA2, GATA6, HNF1A, HOXB13, MAFF, and TCF4 – were screened via Kaplan–Meier curves. Cox analysis determined for the first time that HOXB13 expression and gender were independently associated with LSCC prognosis. Compared to control tissues, elevated expression of HOXB13 was found in LSCC tissues (standardized mean difference = 0.44, 95% confidence interval [0.13–0.76]). HOXB13 expression also makes it feasible to screen LSCC from non-LSCC (area under the curve = 0.77), and HOXB13 may play an essential role in LSCC by regulating HOXB7. In conclusion, HOXB13 may be a novel marker for LSCC clinical screening and treatment.

Revealing the microscopic origin of the unstable nature in organic solar cells (OSCs) is a key issue for further device optimization. Here, we identify that the light-soaking process seen in OSCs could be caused by the effect of space charge polarization in the photoactive layer. The impact of the space charge polarization on the light-soaking is traced experimentally through the determination of dielectric spectra and explored by recombination characterization through charge carrier dynamics study. The Maxwell–Wagner relaxation time is determined from the dielectric spectra. Hence, important information on factors that govern the light-soaking in OSCs is obtained. The results reveal the relationship between dielectric spectra and stability of the devices under illumination, and establish a benchmark for the performance comparison of a set of different cells. Moreover, study of charge carrier dynamics in the device under operation shows that light-soaking and the subsequent space charge polarization result in improved mobility and charge carrier lifetime, thus responsible for the improved device performance. The findings presented here are important for the understanding and optimization of organic solar cells and we expect the presented methods to be generally applicable to other OPV material sets.

van der Waals layered heterojunctions have a variety of band offsets that open up possibilities for a wide range of novel and multifunctional devices. However, due to their poor pristine carrier concentrations and limited band modulation methods, multifunctional p-n heterojunctions are very difficult to achieve. In this report, we developed a highly effective N2O plasma process to treat MoTe2/MoS2 heterojunctions. This allowed us to adjust the hole and electron concentrations in the two materials independently and simultaneously. More importantly, for the first time, we were able to create opposite doping on the two sides of the junction through a single-step treatment. With a very wide doping range from pristine to degenerate levels, a MoTe2/MoS2 heterojunction can be modulated to behave as a forward rectifying diode with enhanced rectifying ratio and as a tunneling transistor with negative differential resistance at room temperature. The new approach provides an effective and generic doping scheme for heterojunctions to construct versatile and multifunctional electronic devices.

Purpose: To evaluate the relationship between corneal biomechanical parameters and lamina cribrosa (LC) curvature in normal tension glaucoma (NTG). Methods: 95 eyes of 56 NTG patients were enrolled in this prospective, observational study. Corneal biomechanical parameters, including stiffness parameters at applanation 1 (SP-A1), deformation amplitude ratio (DA ratio), inverse concave radius and biomechanically corrected intraocular pressure estimate (bIOP), were captured using the Corneal Visualization Scheimpflug Technology instrument (Corvis-ST). LC curvature was evaluated by mean adjusted LC curvature index (maLCCI) averaged by the measurements on 12 radial B-scan images obtained using swept-source optical coherence tomography (SS-OCT). Linear mixed models were constructed to assess the relationship between corneal biomechanical parameters and LC curvature. Results: The mean age of participants was 51.04 ± 13.74 years (range, 24–82 years). The SP-A1 and maLCCI were 93.50 ± 13.82 mm Hg/mm and 7.57 ± 1.58, respectively. In univariate and multivariate analysis, SP-A1 ( p Conclusions: The greater posterior LC curvature was associated with lower corneal stiffness parameters and younger eyes in NTG patients. Clinical Trial Registration: Chinese Clinical Trial Registry, ChiCTR1900021465.

Introduction: To evaluate the association of focal lamina cribrosa (LC) defect with asymmetric visual field (VF) loss in normal-tension glaucoma (NTG) through intereye comparisons. Methods: Paired eyes were divided into better and worse eyes according to the mean deviation (MD), and ocular parameters were compared between them. Furthermore, patients in the asymmetric group were classified as subgroup A (one eye with LC defect and the fellow one without), subgroup B (both eyes without LC defect), and subgroup C (both eyes with LC defect). Generalized estimation equation approach was used to evaluate the association between ocular parameters and asymmetric VF. Results: A total of 140 eyes of 70 NTG patients were included in the asymmetric group. LC defects were more common in better eyes than that in worse eyes (27/70 [38.57%] vs. 10/70 [14.29%], p = 0.001), and all eyes with LC defect had myopia. Multivariate analysis revealed that the presence of LC defect was significantly associated with better eyes in the asymmetric group (odds ratio, 0.27; p = 0.001). For subgroup A, eyes with LC defects exhibited lower peak IOP (p = 0.011) and lower mean IOP (p = 0.018) than the fellow eyes without. In addition, longer AL (p = 0.025) and larger tilt ratio (p = 0.032) were found in eyes with LC defects. For subgroup B without LC defects, larger tilt ratio was shown to be a risk factor for VF loss (odds ratio, 6.13; p = 0.001). There was no significant difference of binocular parameters except for MD (p < 0.001) in subgroup C. Conclusions: LC defects in myopia were suggested to be associated with better eyes in NTG with asymmetric VF loss. However, in patients without LC defect, larger tilt ratio was a risk factor for VF defect. There might be different pathological mechanisms in asymmetric VF loss for different NTG subtypes.

Background Incomplete aneurysmal occlusion is a common feature of immediate posttreatment angiography. The safety and outcomes of acutely ruptured intracranial aneurysms (RIAs) with incomplete occlusion after stent-assisted coiling (SAC) and no-stent coiling (NSC) have not been well clarified. Progressive occlusion of stents can promote the complete occlusion of intracranial aneurysms (IAs), but it remains to be determined if progressive occlusion in acutely RIAs with incomplete occlusion after coiling may be enhanced by protective stenting. This study aimed to evaluate the safety and outcomes of those aneurysms after SAC and NSC; And to discover whether the stents can promote progressive aneurysm occlusion in such lesions or not. Methods We reviewed 199 patients with acutely RIAs underwent endovascular coiling and developed incomplete occlusion in the past seven years. The patients’ clinical and imaging information were recorded and analyzed. Univariate and multivariate analyses were performed to determine the association of recurrence rate with potential risk factors. Results SAC group had wider aneurysms neck (3.471 mm vs 2.830 mm, P = 0.009) and smaller dome-to-neck ratio (1.536 vs 2.111, P = 0.001) than in NSC group. There was no significant difference between the two groups in total procedure-related complications rate (31.7% vs 23.5%, P = 0.195), procedure-related mortality (6.9% vs 2.0%, P = 0.170) and modified Rankin Scale (mRS) score at 6-month follow-up (P > 0.05). However, SAC group had significantly higher ischemic complications rate (21.8% vs 8.2%, P = 0.007) and complete occlusion rate (65.6% vs 48.3%, P = 0.020), and lower recurrence rate (15.6% vs 28.1%, P = 0.042) than NSC group based on 6-month follow-up angiograms. Additionally, Multivariable analysis showed NSC was an independent risk factor for aneurysm recurrence (Odds Ratio [OR]: 4.061; P = 0.018). Conclusions Acutely RIAs with incomplete occlusion after SAC is associated with higher complications rate and mortality, but has an acceptable safety profile and similar clinical outcome compared to NSC, as well as gives patients superior angiography outcome by progressive occlusion of stents.

Artificial intelligence (AI) advancements have significant implications for medical imaging. Stroke is the leading cause of disability and the fifth leading cause of death in the United States. AI applications for stroke imaging are a topic of intense research. AI techniques are well-suited for dealing with vast amounts of stroke imaging data and a large number of multidisciplinary approaches used in classification, risk assessment, segmentation tasks, diagnosis, prognosis, and even prediction of therapy responses. This article addresses this topic and seeks to present an overview of machine learning and/or deep learning applied to stroke imaging.

BACKGROUND: Male Kunming mice were divided into the normal diet group (control group) and the high-fat diet group (HF group, 185 g·kg-1 protein, 600 g·kg-1 fat and 205 g·kg-1 carbohydrate). After eight weeks' feeding, the behavioral and biochemical indicators in serum were determined. The double bottle preference experiment was used to study the preference of mice for five sweeteners. The contents of monoamine neurotransmitters and genes expression related to dopamine (DA) and opioid receptor were also determined. RESULTS: The body weight of the mice in HF group increased significantly (p < 0.05) after 36 days compared with the control group. While the feed intake of HF group increased sharply in the first 12 days, and then it became basically unchanged. The preference of the HF group for all the five sweeteners was very significantly lower (p < 0.01) than that of the control group. The depression-related behavior was observed in the HF group mice. The contents of TG, TC and LDLC in the HF group were extremely higher (p < 0.01) than that of the control group. The genes expression related to DA and opioid receptor in the HF group was significantly lower than that of the control group, except for the preproenkephalin (PENK). CONCLUSIONS: In summary, this study suggested that a long-term high-fat diet could result in the decrease of the preference for sweeteners on mice due to the long-term consumption of high-fat diet could result in a state of reward hypofunction on mice. This article is protected by copyright. All rights reserved.

Flash memories and semiconductor p-n junctions are two elementary but incompatible building blocks of most electronic and optoelectronic devices. The pressing demand to efficiently transfer massive data between memories and logic circuits, as well as for high data storage capability and device integration density, has fueled the rapid growth of technique and material innovations. Two-dimensional (2D) materials are considered as one of the most promising candidates to solve this challenge. However, a key aspect for 2D materials to build functional devices requires effective and accurate control of the carrier polarity, concentration and spatial distribution in the atomically thin structures. Here, a non-volatile opto-electrical doping approach is demonstrated, which enables reversibly writing spatially resolved doping patterns in the MoTe2 conductance channel through a MoTe2/hexagonal boron nitride (h-BN) heterostructure. Based on the doping effect induced by the combination of electrostatic modulation and ultraviolet light illumination, a 3-bit flash memory and various homojunctions on the same MoTe2/BN heterostructure are successfully developed. The flash memory achieved 8 well distinguished memory states with a maximum on/off ratio over 104. Each state showed negligible decay during the retention time of 2,400 s. The heterostructure also allowed the formation of p-p, n-n, p-n, and n-p homojunctions and the free transition among these states. The MoTe2 p-n homojunction with a rectification ratio of 103 exhibited excellent photodetection and photovoltaic performance. Having the memory device and p-n junction built on the same structure makes it possible to bring memory and computational circuit on the same chip, one step further to realize near-memory computing.

Toward the long-standing dream of artificial intelligence, two successful solution paths have been paved: 1) neuromorphic computing and 2) deep learning. Recently, they tend to interact for simultaneously achieving biological plausibility and powerful accuracy. However, models from these two domains have to run on distinct substrates, i.e., neuromorphic platforms and deep learning accelerators, respectively. This architectural incompatibility greatly compromises the modeling flexibility and hinders promising interdisciplinary research. To address this issue, we build a unified model description framework and a unified processing architecture (Tianjic), which covers the full stack from software to hardware. By implementing a set of integration and transformation operations, Tianjic is able to support spiking neural networks, biological dynamic neural networks, multilayered perceptron, convolutional neural networks, recurrent neural networks, and so on. A compatible routing infrastructure enables homogeneous and heterogeneous scalability on a decentralized many-core network. Several optimization methods are incorporated, such as resource and data sharing, near-memory processing, compute/access skipping, and intra-/inter-core pipeline, to improve performance and efficiency. We further design streaming mapping schemes for efficient network deployment with a flexible tradeoff between execution throughput and resource overhead. A 28-nm prototype chip is fabricated with >610-GB/s internal memory bandwidth. A variety of benchmarks are evaluated and compared with GPUs and several existing specialized platforms. In summary, the fully unfolded mapping can achieve significantly higher throughput and power efficiency; the semi-folded mapping can save 30x resources while still presenting comparable performance on average. Finally, two hybrid-paradigm examples, a multimodal unmanned bicycle and a hybrid neural network, are demonstrated to show the potential of our unified architecture. This article paves a new way to explore neural computing.

Despite recent progress on the segmentation of high-resolution images, there exist an unsolved problem, i.e., the trade-off among the segmentation accuracy, memory resources and inference speed. So far, GLNet is introduced for high or ultra-resolution image segmentation, which has reduced the computational memory of the segmentation network. However, it ignores the importances of different cropped patches, and treats tiled patches equally for fusion with the whole image, resulting in high computational cost. To solve this problem, we introduce a patch proposal network (PPN) in this paper, which adaptively distinguishes the critical patches from the trivial ones to fuse with the whole image for refining segmentation. PPN is a classification network which alleviates network training burden and improves segmentation accuracy. We further embed PPN in a global-local segmentation network, instructing global branch and refinement branch to work collaboratively. We implement our method on four image datasets:DeepGlobe, ISIC, CRAG and Cityscapes, the first two are ultra-resolution image datasets and the last two are high-resolution image datasets. The experimental results show that our method achieves almost the best segmentation performance compared with the state-of-the-art segmentation methods and the inference speed is 12.9 fps on DeepGlobe and 10 fps on ISIC. Moreover, we embed PPN with the general semantic segmentation network and the experimental results on Cityscapes which contains more object classes demonstrate the generalization ability on general semantic segmentation.

Recently, due to the availability of big data and the rapid growth of computing power, artificial intelligence (AI) has regained tremendous attention and investment. Machine learning (ML) approaches have been successfully applied to solve many problems in academia and in industry. Although the explosion of big data applications is driving the development of ML, it also imposes severe challenges of data processing speed and scalability on conventional computer systems. Computing platforms that are dedicatedly designed for AI applications have been considered, ranging from a complement to von Neumann platforms to a “must-have” and stand-alone technical solution. These platforms, which belong to a larger category named “domain-specific computing,” focus on specific customization for AI. In this article, we focus on summarizing the recent advances in accelerator designs for deep neural networks (DNNs)—that is, DNN accelerators. We discuss various architectures that support DNN executions in terms of computing units, dataflow optimization, targeted network topologies, architectures on emerging technologies, and accelerators for emerging applications. We also provide our visions on the future trend of AI chip designs. Keywords: Deep neural network, Domain-specific architecture, Accelerator

Energy loss within organic solar cells (OSCs) is undesirable as it reduces cell efficiency1–4. In particular, non-radiative recombination loss3 and energetic disorder5, which are closely related to the tail states below the band edge and the overall photon energy loss, need to be minimized to improve cell performance. Here, we report how the use of a small-molecule acceptor with torsion-free molecular conformation can achieve a very low degree of energetic disorder and mitigate energy loss in OSCs. The resulting single-junction OSC has an energy loss due to non-radiative recombination of just 0.17 eV and a high power conversion efficiency of up to 16.54% (certified as 15.89% by the National Renewable Energy Laboratory). The findings take studies of organic photovoltaics deeper into a new regime, beyond the limits of energetic disorder and large energy offset for charge generation. An organic solar cell designed with minimal energetic disorder exhibits very low energy loss due to non-radiative recombination and highly efficient operation.

Mounting evidence indicates the impact of gender difference on the assessment, treatment, and outcome of patients with acute myocardial infarction (AMI). However, gender-specific prognostic markers of AMI are still lacking. The present study aimed to investigate gender-specific markers of poor prognosis (all-cause mortality or readmission) in a cohort of AMI patients followed up for 6 months. Compared with males (n = 157), females (n = 40) were older and more frequent with previous medical history of hypertension and diabetes mellitus. During the 6-month follow-up, BUN ≥ 7.73 mM, myoglobin ≥ 705.8 ng/mL, and Killip classification 2-4 were identified as the independent predictors of poor prognosis for male AMI patients, while D-dimer ≥ 0.43 mg/L as an independent predictor of poor prognosis in female AMI patients. In conclusion, our data suggest that prognostic markers for AMI patients may differ according to genders. Gender-specific prognostic markers may be useful to guide the risk stratification, clinical therapy, and medications for AMI patients.

Existing successful person re-identification (Re-ID) models often employ the part-level representation to extract the fine-grained information, but commonly use the loss that is particularly designed for global features, ignoring the relationship between semantic parts. In this paper, we present a novel triplet loss that emphasizes the salient parts and also takes the consideration of alignment. This loss is based on the crossing-bing matching metric that also known as Wasserstein Distance. It measures how much effort it would take to move the embeddings of local features to align two distributions, such that it is able to find an optimal transport matrix to re-weight the distance of different local parts. The distributions in support of local parts is produced via a new attention mechanism, which is calculated by the inner product between high-level global feature and local features, representing the importance of different semantic parts w.r.t. identification. We show that the obtained optimal transport matrix can not only distinguish the relevant and misleading parts, and hence assign different weights to them, but also rectify the original distance according to the learned distributions, resulting in an elegant solution for the mis-alignment issue. Besides, the proposed method is easily implemented in most Re-ID learning system with end-to-end training style, and can obviously improve their performance. Extensive experiments and comparisons with recent Re-ID methods manifest the competitive performance of our method.

Charge-trapping memory devices based on two-dimensional (2D) material heterostructures possess an atomically thin structure and excellent charge transport capability, making them promising candidates for next-generation flash memories to achieve miniaturized size, high storage capacity, fast switch speed, and low power consumption. Here, we report a nonvolatile floating-gate memory device based on an ReS2/boron nitride/graphene heterostructure. The implemented ReS2 memory device displays a large memory window exceeding 100 V, leading to an ultrahigh current ratio over 108 between programming and erasing states. The ReS2 memory device also exhibits an ultrafast switch speed of 1 μs. In addition, the device can endure hundreds of switching cycles and shows stable retention characteristics with ∼40% charge remaining after 10 years. More importantly, taking advantage of its anisotropic electrical properties, a single ReS2 flake can achieve direction-sensitive multi-level data storage to enhance the data storage density. On the basis of these characteristics, the proposed ReS2 memory device is potentially able to serve the entire memory device hierarchy, meeting the need for scalability, capacity, speed, retention, and endurance at each level.

This article presents a tensor multi-task model for person re-identification (Re-ID). Due to discrepancy among cameras, our approach regards Re-ID from multiple cameras as different but related classification tasks, each task corresponding to a specific camera. In each task, we distinguish the person identity as a one-vs-all linear classification problem, where one classifier is associated with a specific person. By constructing all classifiers into a task-specific projection matrix, the proposed method could utilize all the matrices to form a tensor structure, and jointly train all the tasks in a uniform tensor space. In this space, by assuming the features of the same person under different cameras are generated from a latent subspace, and different identities under the same perspective share similar patterns, the high-order correlations, not only across different tasks but also within a certain task, can be captured by utilizing a new type of low-rank tensor constraint. Therefore, the learned classifiers transform the original feature vector into the latent space, where feature distributions across cameras can be well-aligned. Moreover, this model can be incorporated into multiple visual features to boost the performance, and easily extended to the unsupervised setting. Extensive experiments and comparisons with recent Re-ID methods manifest the competitive performance of our method.

Artificial neural networks (ANNs), a popular path towards artificial intelligence, have experienced remarkable success via mature models, various benchmarks, open-source datasets, and powerful computing platforms. Spiking neural networks (SNNs), a category of promising models to mimic the neuronal dynamics of the brain, have gained much attention for brain inspired computing and been widely deployed on neuromorphic devices. However, for a long time, there are ongoing debates and skepticisms about the value of SNNs in practical applications. Except for the low power attribute benefit from the spike-driven processing, SNNs usually perform worse than ANNs especially in terms of the application accuracy. Recently, researchers attempt to address this issue by borrowing learning methodologies from ANNs, such as backpropagation, to train high-accuracy SNN models. The rapid progress in this domain continuously produces amazing results with ever-increasing network size, whose growing path seems similar to the development of deep learning. Although these ways endow SNNs the capability to approach the accuracy of ANNs, the natural superiorities of SNNs and the way to outperform ANNs are potentially lost due to the use of ANN-oriented workloads and simplistic evaluation metrics. In this paper, we take the visual recognition task as a case study to answer the questions of "what workloads are ideal for SNNs and how to evaluate SNNs makes sense". We design a series of contrast tests using different types of datasets (ANN-oriented and SNN-oriented), diverse processing models, signal conversion methods, and learning algorithms. We propose comprehensive metrics on the application accuracy and the cost of memory & compute to evaluate these models, and conduct extensive experiments. We evidence the fact that on ANN-oriented workloads, SNNs fail to beat their ANN counterparts; while on SNN-oriented workloads, SNNs can fully perform better. We further demonstrate that in SNNs there exists a trade-off between the application accuracy and the execution cost, which will be affected by the simulation time window and firing threshold. Based on these abundant analyses, we recommend the most suitable model for each scenario. To the best of our knowledge, this is the first work using systematical comparisons to explicitly reveal that the straightforward workload porting from ANNs to SNNs is unwise although many works are doing so and a comprehensive evaluation indeed matters. Finally, we highlight the urgent need to build a benchmarking framework for SNNs with broader tasks, datasets, and metrics.

Convolution neural networks (CNNs) have demonstrated superior capability in computer vision, speech recognition, autonomous driving, and so forth, which are opening up an artificial intelligence (AI) era. However, conventional CNNs require significant matrix computation and memory usage leading to power and memory issues for mobile deployment and embedded chips. On the algorithm side, the emerging binary neural networks (BNNs) promise portable intelligence by replacing the costly massive floating-point compute-and-accumulate operations with lightweight bit-wise XNOR and popcount operations. On the hardware side, the computing-in-memory (CIM) architectures developed by the non-volatile memory (NVM) present outstanding performance regarding high speed and good power efficiency. In this paper, we propose an NVM-based CIM architecture employing a Preset-XNOR operation in/with the spin–orbit torque magnetic random access memory (SOT-MRAM) to accelerate the computation of BNNs (PXNOR-BNN). PXNOR-BNN performs the XNOR operation of BNNs inside the computing-buffer array with only slight modifications of the peripheral circuits. Based on the layer evaluation results, PXNOR-BNN can achieve similar performance compared with the read-based SOT-MRAM counterpart. Finally, the end-to-end estimation demonstrates $12.3\times $ speedup compared with the baseline with 96.6-image/s/W throughput efficiency.

To compare the clinical results and functional outcomes between two-dimensional conventional radiation therapy (2DRT) and intensity-modulated radiation therapy (IMRT) in nasopharyngeal carcinoma (NPC) with skull-base invasion. A total of 1258 patients were subclassified into two groups: mild skull-base invasion group (792; 63%) and severe skull-base invasion group (466; 37%). Patients were pair matched (1:1 ratio) using six clinical factors into 2DRT or IMRT groups. The Kaplan–Meier method and Cox regression model were performed to assess overall survival (OS), disease-free survival (DFS), locoregional relapse-free survival (LRRFS) and distant metastasis-free survival (DMFS). Toxicities were evaluated. IMRT significantly improved four-year OS compared with 2DRT (65.6% vs. 81.8%, P = 0.000), DFS (57.3% vs. 73.3%, P = 0.000) and LRRFS (76.5% vs. 87.5%, P = 0.003) in NPC with severe skull-base invasion, but similar results were observed in patients with mild skull-base invasion (P > 0.05). In patients with severe invasion, radiation therapy techniques were found to be an independent prognostic factor for OS (HR = 0.457, P = 0.000), DFS (HR = 0.547, P = 0.000) and LRRFS (HR = 0.503, P = 0.004). IMRT was associated with better OS. In subgroups analysis, IMRT group also had a better survival in OS, DFS (P

Deep learning (DL) accelerators suffer from a new security problem, i.e., being vulnerable to physical access based attacks. An adversary can easily obtain the entire neural network (NN) model by physically snooping the memory bus that connects the accelerator chip with DRAM memory. Therefore, memory encryption becomes important for DL accelerators to improve their security. Nevertheless, we observe that traditional memory encryption techniques that have been efficiently used in CPU systems cause significant performance degradation when directly used in DL accelerators, due to the big bandwidth gap between the memory bus and the encryption engine. To address this problem, our paper proposes SEAL, a Secure and Efficient Accelerator scheme for deep Learning to enhance the performance of encrypted DL accelerators by improving the data access bandwidth. Specifically, SEAL leverages a criticality-aware smart encryption scheme that identifies partial data having no impact on the security of NN models and allows them to bypass the encryption engine, thus reducing the amount of data to be encrypted without affecting security. Extensive experimental results demonstrate that, compared with existing memory encryption techniques, SEAL achieves 1.34 – 1.4× overall performance improvement.

Liver has an ability to regenerate itself in mammals, whereas the mechanism has not been fully explained. Here we used a GC/MS-based metabolomic method to profile the dynamic endogenous metabolic change in the serum of C57BL/6J mice at different times after 2/3 partial hepatectomy (PHx), and nine machine learning methods including Least Absolute Shrinkage and Selection Operator Regression (LASSO), Partial Least Squares Regression (PLS), Principal Components Regression (PCR), k-Nearest Neighbors (KNN), Support Vector Machines (SVM), Random Forest (RF), eXtreme Gradient Boosting (xgbDART), Neural Network (NNET) and Bayesian Regularized Neural Network (BRNN) were used for regression between the liver index and metabolomic data at different stages of liver regeneration. We found a tree-based random forest method that had the minimum average Mean Absolute Error (MAE), Root Mean Squared Error (RMSE) and the maximum R square (R2) and is time-saving. Furthermore, variable of importance in the project (VIP) analysis of RF method was performed and metabolites with VIP ranked top 20 were selected as the most critical metabolites contributing to the model. Ornithine, phenylalanine, 2-hydroxybutyric acid, lysine, etc. were chosen as the most important metabolites which had strong correlations with the liver index. Further pathway analysis found Arginine biosynthesis, Pantothenate and CoA biosynthesis, Galactose metabolism, Valine, leucine and isoleucine degradation were the most influenced pathways. In summary, several amino acid metabolic pathways and glucose metabolism pathway were dynamically changed during liver regeneration. The RF method showed advantages for predicting the liver index after PHx over other machine learning methods used and a metabolic clock containing four metabolites is established to predict the liver index during liver regeneration.

Extreme classification (XC) is the essential component of large-scale Deep Learning Systems for a wide range of application domains, including image recognition, language modeling, and recommendation. As classification categories keep scaling in real-world applications, the classifier’s parameters could reach several thousands of Gigabytes, way exceed the on-chip memory capacity. With the advent of near-memory processing (NMP) architectures, offloading the XC component onto NMP units could alleviate the memory-intensive problem. However, naive NMP design with limited area and power budget cannot afford the computational complexity of full classification. To tackle the problem, we first propose a novel screening method to reduce the computation and memory consumption by efficiently approximating the classification output and identifying a small portion of key candidates that require accurate results. Then, we design a new extreme-classification-tailored NMP architecture, namely ENMC, to support both screening and candidates-only classification. Overall, our approximate screening method achieves 7.3 × speedup over the CPU baseline, and ENMC further improves the performance by 7.4 × and demonstrates 2.7 × speedup compared with the state-of-the-art NMP baseline.

Streaming graphs are ubiquitous in today’s big data era. Prior work has improved the performance of streaming graph workloads without taking input characteristics into account. In this work, we demonstrate that input knowledge-driven software and hardware co-design is critical to optimize the performance of streaming graph processing. To improve graph update efficiency, we first characterize the performance trade-offs of input-oblivious batch reordering. Guided by our findings, we propose input-aware batch reordering to adaptively reorder input batches based on their degree distributions. To complement adaptive batch reordering, we propose updating graphs dynamically, based on their input characteristics, either in software (via update search coalescing) or in hardware (via acceleration support). To improve graph computation efficiency, we present input-aware work aggregation which adaptively modulates the computation granularity based on inter-batch locality characteristics. Evaluated across 260 workloads, our input-aware techniques provide on average 4.55 × and 2.6 × improvement in graph update performance for different input types (on top of eliminating the performance degradation from input-oblivious batch reordering). The graph compute performance is improved by 1.26 × (up to 2.7 ×).

Despite recent progress on semantic segmentation, there still exist huge challenges in high or ultra-high resolution images semantic segmentation. Although the latest collaborative global-local semantic segmentation methods such as GLNet [4] and PPN [18] have achieved impressive results, they are inefficient and not fit for practical applications. Thus, in this paper, we propose a novel and efficient collaborative global-local framework on the basis of PPN named Faster-PPN for high or ultra-high resolution images semantic segmentation which makes a better trade-off between the efficient and effectiveness towards the real-time speed. Specially, we propose Dual Mutual Learning to improve the feature representation of global and local branches, which conducts knowledge distillation mutually between the global and local branches. Furthermore, we design the Pixel Proposal Fusion Module to conduct the fine-grained selection mechanism which further reduces the redundant pixels for fusion resulting in the improvement of inference speed. The experimental results on three challenging high or ultra-high resolution datasets DeepGlobe, ISIC and BACH demonstrate that Faster-PPN achieves the best performance on accuracy, inference speed and memory usage compared with state-of-the-art approaches. Especially, our method achieves real-time and near real-time speed with 36 FPS and 17.7 FPS on ISIC and DeepGlobe, respectively.

Named Entity Recognition (NER) is a key task in Natural Language Processing (NLP). In medical domain, NER is very important phase in all end-to-end systems. In this paper, we investigate the performance of NER for disease (D-NER). TaggerOne was evaluated on 52 cardiovascular-related clinical case reports against hand annotation for diseases. Different training sets have been used to evaluate the performance of TaggerOne as a famous tool for NER in biomedical domain.

A quantum computing system naturally consists of two components, the software system and the hardware system. Quantum applications are programmed using the quantum software and then executed on the quantum hardware. However, the performance of existing quantum computing system is still limited. Solving a practical problem that is beyond the capability of classical computers on a quantum computer has not yet been demonstrated. In this review, we point out that the quantum software and hardware systems should be designed collaboratively to fully exploit the potential of quantum computing. We first review three related works, including one hardware-aware quantum compiler optimization, one application-aware quantum hardware architecture design flow, and one co-design approach for the emerging quantum computational chemistry. Then we discuss some potential future directions following the co-design principle.

Background and AimsThe heat shock protein (Hsp) 90α is induced by stress and regulates inflammation through multiple pathways. Elevated serum Hsp90α had been found in nonalcoholic steatohepatitis (NASH). Geranylgeranylacetone (GGA, also called teprenone) is a terpenoid derivative. It was reported to induce Hsp and alleviate insulin resistance. We aimed to evaluate the Hsp90α as a biomarker in predicting metabolic-associated fatty liver disease (MAFLD) and define the therapeutic effects of geranylgeranylacetone for the disease.MethodsA clinical study was conducted to analyze the elements associated with Hsp90α, and a predictive model of MAFLD was developed based on Hsp90α. The histopathological correlation between Hsp90α and MAFLD was investigated through a diet-induced mouse model. Furthermore, GGA was applied to the mouse model.ResultsSerum Hsp90α was increased in patients with MAFLD. A positive linear relationship was found between age, glycosylated hemoglobin (HbA1c), MAFLD, and serum Hsp90α. Meanwhile, a negative linear relationship with body mass index (BMI) was found. A model using Hsp90α, BMI, HbA1c, and ALT was established for predicting MAFLD. The area under the receiver operating characteristic (ROC) curves was 0.94 (95% CI 0.909–0.971, p = 0.000). The sensitivity was 84.1%, and the specificity was 93.1%. In vitro experiments, GGA induced Hsp90α in steatosis cells. In the mice model, Hsp90α decreased in the GGA treatment group. Hepatic steatosis, inflammation, insulin resistance, and glucose intolerance were improved in the GGA-treated group. Serum Hsp90α was positively correlated with steatohepatitis activity according to hepatic histopathology.ConclusionsSerum Hsp90α was elevated in MAFLD, and a positive correlation between serum Hsp90α and the grade of activity of steatohepatitis was observed. The model using BMI, HbA1c, and alanine aminotransferase (ALT) had a good value to predict MAFLD. The findings also revealed the effectiveness of GGA in the treatment of MAFLD.

Face anti-spoofing approaches based on domain generalization (DG) have drawn growing attention due to their robustness for unseen scenarios. Previous methods treat each sample from multiple domains indiscriminately during the training process, and endeavor to extract a common feature space to improve the generalization. However, due to complex and biased data distribution, directly treating them equally will corrupt the generalization ability. To settle the issue, we propose a novel Dual Reweighting Domain Generalization (DRDG) framework which iteratively reweights the relative importance between samples to further improve the generalization. Concretely, Sample Reweighting Module is first proposed to identify samples with relatively large domain bias, and reduce their impact on the overall optimization. Afterwards, Feature Reweighting Module is introduced to focus on these samples and extract more domain-irrelevant features via a self-distilling mechanism. Combined with the domain discriminator, the iteration of the two modules promotes the extraction of generalized features. Extensive experiments and visualizations are presented to demonstrate the effectiveness and interpretability of our method against the state-of-the-art competitors.

Human beings have a great generalization ability to recognize a novel category by only seeing a few number of samples. This is because humans possess the ability to learn from the concepts that already exist in our minds. However, many existing few-shot approaches fail in addressing such a fundamental problem, {\it i.e.,} how to utilize the knowledge learned in the past to improve the prediction for the new task. In this paper, we present a novel purified memory mechanism that simulates the recognition process of human beings. This new memory updating scheme enables the model to purify the information from semantic labels and progressively learn consistent, stable, and expressive concepts when episodes are trained one by one. On its basis, a Graph Augmentation Module (GAM) is introduced to aggregate these concepts and knowledge learned from new tasks via a graph neural network, making the prediction more accurate. Generally, our approach is model-agnostic and computing efficient with negligible memory cost. Extensive experiments performed on several benchmarks demonstrate the proposed method can consistently outperform a vast number of state-of-the-art few-shot learning methods.

Background: Recent studies have demonstrated that airway basal stem cells (BCs) transplantation can ameliorate bleomycin-induced idiopathic pulmonary fibrosis (IPF) through lung regeneration promotion. However, BCs under oxidative stress in the alveolar microenvironment are poor in survival, causing unsatisfied efficacy of BCs transplantation. In this study, we investigated whether Coenzyme Q10(CoQ10) counteracts oxidative stress in the alveolar microenvironment, thus improved the efficacy of BCs transplantation for IPF treatment.Methods: The protective effects of CoQ10 on H2O2-induced BCs apoptosis and cytoplasmic reactive oxygen species (ROS) level were tested by flow cytometry in vitro. The therapeutic effects of BCs combined with CoQ10 were compared to a single BCs transplantation protocol in IPF treatment after two weeks and were evaluated by parameters including changes of body weight and survival rate, as well as various levels of pulmonary inflammation, α-SMA expression and hydroxyproline (HYP) in IPF mice lung tissues.Results: CoQ 10 preincubation with BCs (10 mM, 24 h) significantly reduced the late apoptosis of BCs and the number of oxidative stressful BCs as a result of H2O2 stimulation (1mM, 6h) in vitro. IPF mice models were constructed through bleomycin (5 mg/Kg) intratracheal instillation. Bleomycin-induced IPF mice showed weight loss continuously and mortality increased progressively during modeling. Serious pulmonary inflammatory cell infiltration, collagen fiber proliferation, and collagen protein deposition were observed in lung tissues of IPF mice. Though BCs transplantation alone improved indicators above in bleomycin-induced IPF mice to some extent, the combination with CoQ10 improved the transplantation efficacy and obtained better therapeutic effects.Conclusion：CoQ10 blocked H2O2-induced apoptosis of BCs and ROS production in vitro, and enhanced the efficacy of BCs transplantation on bleomycin-induced IPF in mice.

Separation of heart sound signals from the recorded mixtures has become a hot research topic in the clinical diagnosis of heart diseases. To better understand the current health status of patients, it is of great significance to develop intelligent auscultation methods to improve the effectiveness of hearing and assist clinicians in practice. In this paper, an optimization convolutive blind source separation algorithm utilizing independent vector analysis is proposed for separation of the heart sound mixtures. In the algorithm, a denoising approach is firstly used to reduce the impact of the additive white Gaussian noise. Then, using the short time Fourier transform, the mixing signals in time domain are transmitted into the frequency domain. Afterwards, the unmixng matrix is updated using Newton’s method, and the heart sound sources are reconstructed based on the estimated unmixing matrix. Experimental results show that the algorithm obtains better separation performance than the baseline methods. Especially, it has better superiority in the strong reverberant environment.

Few-shot learning (FSL), aiming to extract new knowledge from very small amount of labeled samples, has attracted noticeable attentions recently. However, most of existing methods often fail when facing huge domain shift between seen and unseen classes. We think this should be attributed to the episode strategy which ignore utilizing support samples to induct the test classes. So in this paper, for the first time, we propose a bilevel episode strategy (BL-ES) to train a inductive graph network (IGN) that learn to both comparison and induction. Specifically, first, outer episodes in BL-ES simulate the cross-domain few-shot tasks constantly, while inner episodes learn to drive IGN to induct the common features of test classes. Then, the propsoed IGN captures the correlation among all samples to update meta points of each category in induction module. Finally, we introduce a geometrical constraint term utilizing meta points into the training loss, to update the nodes and edges in feature space. This way improves the robustness of training process. Extensive experiments show that our framework outperforms the state-of-the-art FSL alternatives, and are more suitable for real-world applications.