Analysis of Risk Factors for Death in the Coronavirus Disease 2019 (COVID-19) Population: Data Analysis from a Large General Hospital in Anhui, China.

During the coronavirus disease 2019 (COVID-19) pandemic, clinical prevention, early diagnosis, and hematological monitoring were challenging areas. This study aims to compare risk factors and hematological and biochemical data in non-survivor group patients with COVID-19 versus survivor group patients. A total of 204 patients with COVID-19 were selected as research subjects from December 2022 to January 2023. We analyzed the age, sex, time from onset to admission, and laboratory test indicators upon admission. The differences between surviving and deceased patients and mortality-related risk factors were examined. Among the 204 patients, 168 survived, whereas 36 died during hospitalization. Significant differences were observed between the two groups with COVID-19 across various factors, including age (p < 0.0001), WBC count (p < 0.0001), RBC count (p < 0.05), neutrophils (p < 0.0001), lymphocytes (p < 0.05), mean corpuscular hemoglobin concentration (MCHC) (p < 0.0001), RBC distribution width-standard deviation (RDW-SD) (p < 0.0001), RBC distribution width coefficient of variation (RDW-CV) (p < 0.0001), aspartate aminotransferase (AST) (p < 0.05), albumin (ALB) (p < 0.0001), creatinine (CR) (p < 0.0001), uric acid (UA) (p < 0.0001), blood urea nitrogen (BUN) (p < 0.0001), plasma thrombin time (TT) (p < 0.05), prothrombin time (PT) (p < 0.0001), and D-dimer (p < 0.0001). Multivariate logistic analysis revealed that older age, CR, UA, and ALB were independent factors associated with death (p < 0.05). Elderly patients with underlying diseases, abnormal routine blood test indices, and abnormal renal function and coagulation indices are at an increased worse prognosis and should be identified early. Age, UA, CR, and ALB can be used as predictors to assess the worse prognosis in the hospital.

Recent advances in early pregnancy loss diagnosis in dairy cows: New approaches.

Early pregnancy loss is a primary cause of low reproductive rates in dairy cows, posing severe economic losses to dairy farming. The accurate diagnosis of dairy cows with early pregnancy loss allows for oestrus synchronization, shortening day open, and increasing the overall conception rate of the herd. Several techniques are available for detecting early pregnancy loss in dairy cows, including rectal ultrasound, circulating blood progesterone, and pregnancy-associated glycoproteins (PAGs). Yet, there is a need to improve on existing techniques and develop novel strategies to identify cows with early pregnancy loss accurately. This manuscript reviews the applications of rectal ultrasound, circulating blood progesterone concentration, and PAGs in the diagnosis of pregnancy loss in dairy cows. The manuscript also discusses the recent progress of new technologies, including colour Doppler ultrasound (CDUS), interferon tau-induced genes (ISGs), and exosomal miRNA in diagnosing pregnancy loss in dairy cows. This study will provide an option for producers to re-breed cows with pregnancy loss, thereby reducing the calving interval and economic costs. Meanwhile, this manuscript might also act as a reference for exploring more economical and precise diagnostic technologies for early pregnancy loss in dairy cows.

Genetic associations of protein-coding variants in venous thromboembolism.

Previous genetic studies of venous thromboembolism (VTE) have been largely limited to common variants, leaving the genetic determinants relatively incomplete. We performed an exome-wide association study of VTE among 14,723 cases and 334,315 controls. Fourteen known and four novel genes (SRSF6, PHPT1, CGN, and MAP3K2) were identified through protein-coding variants, with broad replication in the FinnGen cohort. Most genes we discovered exhibited the potential to predict future VTE events in longitudinal analysis. Notably, we provide evidence for the additive contribution of rare coding variants to known genome-wide polygenic risk in shaping VTE risk. The identified genes were enriched in pathways affecting coagulation and platelet activation, along with liver-specific expression. The pleiotropic effects of these genes indicated the potential involvement of coagulation factors, blood cell traits, liver function, and immunometabolic processes in VTE pathogenesis. In conclusion, our study unveils the valuable contribution of protein-coding variants in VTE etiology and sheds new light on its risk stratification.

A Versatile Framework for Multi-Scene Person Re-Identification.

Person Re-identification (ReID) has been extensively developed for a decade in order to learn the association of images of the same person across non-overlapping camera views. To overcome significant variations between images across camera views, mountains of variants of ReID models were developed for solving a number of challenges, such as resolution change, clothing change, occlusion, modality change, and so on. Despite the impressive performance of many ReID variants, these variants typically function distinctly and cannot be applied to other challenges. To our best knowledge, there is no versatile ReID model that can handle various ReID challenges at the same time. This work contributes to the first attempt at learning a versatile ReID model to solve such a problem. Our main idea is to form a two-stage prompt-based twin modeling framework called VersReID. Our VersReID firstly leverages the scene label to train a ReID Bank that contains abundant knowledge for handling various scenes, where several groups of scene-specific prompts are used to encode different scene-specific knowledge. In the second stage, we distill a V-Branch model with versatile prompts from the ReID Bank for adaptively solving the ReID of different scenes, eliminating the demand for scene labels during the inference stage. To facilitate training VersReID, we further introduce the multi-scene properties into self-supervised learning of ReID via a multi-scene prioris data augmentation (MPDA) strategy. Through extensive experiments, we demonstrate the success of learning an effective and versatile ReID model for handling ReID tasks under multi-scene conditions without manual assignment of scene labels in the inference stage, including general, low-resolution, clothing change, occlusion, and cross-modality scenes. Codes and models will be made publicly available.

A fresh perspective on dissociation mechanism of cellulose in DMAc/LiCl system based on Li bond theory.

In this case, various characterization technologies have been employed to probe dissociation mechanism of cellulose in N,N-dimethylacetamide/lithium chloride (DMAc/LiCl) system. These results indicate that coordination of DMAc ligands to the Li+-Cl- ion pair results in the formation of a series of Lix(DMAc)yClz (x = 1, 2; y = 1, 2, 3, 4; z = 1, 2) complexes. Analysis of interaction between DMAc ligand and Li center indicate that Li bond plays a major role for the formation of these Lix(DMAc)yClz complexes. And the saturation and directionality of Li bond in these Lix(DMAc)yClz complexes are found to be a tetrahedral structure. The hydrogen bonds between two cellulose chains could be broken at the nonreduced end of cellulose molecule via combined effects of basicity of Cl- ion and steric hindrance of [Li (DMAc)4]+ unit. The unique feature of Li bond in Lix(DMAc)yClz complexes is a key factor in determination of the dissociation mechanism.

Large-scale whole-exome sequencing of neuropsychiatric diseases and traits in 350,770 adults.

While numerous genomic loci have been identified for neuropsychiatric conditions, the contribution of protein-coding variants has yet to be determined. Here we conducted a large-scale whole-exome-sequencing study to interrogate the impact of protein-coding variants on 46 neuropsychiatric diseases and 23 traits in 350,770 adults from the UK Biobank. Twenty new genes were associated with neuropsychiatric diseases through coding variants, among which 16 genes had impacts on the longitudinal risks of diseases. Thirty new genes were associated with neuropsychiatric traits, with SYNGAP1 showing pleiotropic effects across cognitive function domains. Pairwise estimation of genetic correlations at the coding-variant level highlighted shared genetic associations among pairs of neurodegenerative diseases and mental disorders. Lastly, a comprehensive multi-omics analysis suggested that alterations in brain structures, blood proteins and inflammation potentially contribute to the gene-phenotype linkages. Overall, our findings characterized a compendium of protein-coding variants for future research on the biology and therapeutics of neuropsychiatric phenotypes.

In-house ammonia induced lung impairment and oxidative stress of ducks.

Ammonia (NH3) is a toxic gas that in intensive poultry houses, damages the poultry health and induces various diseases. This study investigated the effects of NH3 exposure (0, 15, 30, and 45 ppm) on growth performance, serum biochemical indexes, antioxidative indicators, tracheal and lung impairments in Pekin ducks. A total of 288 one-day-old Pekin male ducks were randomly allocated to 4 groups with 6 replicates and slaughtered after the 21-d test period. Our results showed that 45 ppm NH3 significantly reduced the average daily feed intake (ADFI) of Pekin ducks. Ammonia exposure significantly reduced liver, lung, kidney, and heart indexes, and lowered the relative weight of the ileum. With the increasing of in-house NH3, serum NH3 and uric acid (UA) concentrations of ducks were significantly increased, as well as liver malondialdehyde (MDA), superoxide dismutase (SOD), and glutathione peroxidase (GPX-Px) contents. High NH3 also induced trachea and lung injury, thereby increasing levels of tumor necrosis factor-α (TNF-α) and interleukin-4 (IL-4) in the lung, and decreasing the mRNA expressions of zonula occludens 1 (ZO-1) and claudin 3 (CLDN3) in the lung. In conclusion, in-house NH3 decrease the growth performance in ducks, induce trachea and lung injuries and meanwhile increase the compensatory antioxidant activity for host protection.

Comparison of PD-L1 (22C3) Expression in Paired Primary and Metastatic Breast Carcinoma.

INTRODUCTION: PD-L1 immunohistochemistry (IHC) is being used as a predictive marker of the benefit derived from immunotherapy in several cancer types, including breast cancer. However, the insight gleaned of the prognostic and predictive value of PD-L1 status and its correlation with molecular characteristics during breast cancer progression remains limited. METHODS: We performed an PD-L1 (22C3) assay in pre-treatment primary and metastatic tumor sections from 33 patients with breast carcinoma, matched for post neoadjuvant chemotherapy (p-NACT). PD-L1 expression was evaluated using 3 scoring methods: immune cell (IC) and tumor cell (TC) with a 1% as the cutoff value, and combined positive scores (CPS) with a 1 as the cutoff value. Twenty-two samples from 11 patients had successful fluorescence in situ hybridization (FISH)-based molecular data available for analysis. RESULTS: In the 33 pre-treatment primary tumors, PD-L1 IC, TC, and CPS showed positive correlation with stromal tumor infiltrate lymphocytes (sTIL), histological grade 3, and triple negative breast carcinoma (TNBC). In the matched metastatic tumors, only PD-L1 IC showed a positive correlation with sTIL. The primary tumors showed a higher PD-L1 expression than the matched metastatic tumors by IC and CPS. Negative to positive conversion by CPS was identified in the metastatic tumors from lung, pleura and liver. p-NACT tumors also showed a trend of lower PD-L1 expression compared to the pre-treatment tumors. Six patients had matched samples for molecular and PD-L1 comparison, and none of them showed consistent gene alterations or PD-L1 expression among the primary, p-NACT and metastatic tumors. CONCLUSION: Our study showed a decrease in PD-L1 expression and disconnected molecular features during breast cancer progression. Repeating PD-L1 IHC testing could be considered in some specific metastatic sites if primary tumors were negative. Further studies are needed to identify other predictive factors for immune checkpoint inhibitor (ICI) therapy in patients with breast carcinoma.

Effects of opioid-free anaesthesia compared with balanced general anaesthesia on nausea and vomiting after video-assisted thoracoscopic surgery: a single-centre randomised controlled trial.

OBJECTIVES: Opioid-free anaesthesia (OFA) has emerged as a promising approach for mitigating the adverse effects associated with opioids. The objective of this study was to evaluate the impact of OFA on postoperative nausea and vomiting (PONV) following video-assisted thoracic surgery. DESIGN: Single-centre randomised controlled trial. SETTING: Tertiary hospital in Shanghai, China. PARTICIPANTS: Patients undergoing video-assisted thoracic surgery were recruited from September 2021 to June 2022. INTERVENTION: Patients were randomly allocated to OFA or traditional general anaesthesia with a 1:1 allocation ratio. PRIMARY AND SECONDARY OUTCOME MEASURES: The primary outcome measure was the incidence of PONV within 48 hours post-surgery, and the secondary outcomes included PONV severity, postoperative pain, haemodynamic changes during anaesthesia, and length of stay (LOS) in the recovery ward and hospital. RESULTS: A total of 86 and 88 patients were included in the OFA and control groups, respectively. Two patients were excluded because of severe adverse events including extreme bradycardia and epilepsy-like convulsion. The incidence and severity of PONV did not significantly differ between the two groups (29 patients (33.0%) in the control group and 22 patients (25.6%) in the OFA group; relative risk 0.78, 95% CI 0.49 to 1.23; p=0.285). Notably, the OFA approach used was associated with an increase in heart rate (89±17 vs 77±15 beats/min, t-test: p<0.001; U test: p<0.001) and diastolic blood pressure (87±17 vs 80±13 mm Hg, t-test: p=0.003; U test: p=0.004) after trachea intubation. Conversely, the control group exhibited more median hypotensive events per patient (mean 0.5±0.8 vs 1.0±2.0, t-test: p=0.02; median 0 (0-4) vs 0 (0-15), U test: p=0.02) during surgery. Postoperative pain scores, and LOS in the recovery ward and hospital did not significantly differ between the two groups. CONCLUSIONS: Our study findings suggest that the implementation of OFA does not effectively reduce the incidence of PONV following thoracic surgery when compared with traditional total intravenous anaesthesia. The opioid-free strategy used in our study may be associated with severe adverse cardiovascular events. TRIAL REGISTRATION NUMBER: ChiCTR2100050738.

IL-6 mediates olfactory dysfunction in a mouse model of allergic rhinitis.

BACKGROUND: Immune-inflammatory response is a key element in the occurrence and development of olfactory dysfunction (OD) in patients with allergic rhinitis (AR). As one of the core factors in immune-inflammatory responses, interleukin (IL)-6 is closely related to the pathogenesis of allergic diseases. It may also play an important role in OD induced by diseases, such as Sjögren's syndrome and coronavirus disease 2019. However, there is no study has reported its role in OD in AR. Thus, this study aimed to investigate the role of IL-6 in AR-related OD, in an attempt to discover a new target for the prevention and treatment of OD in patients with AR. METHODS: Differential expression analysis was performed using the public datasets GSE52804 and GSE140454 for AR, and differentially expressed genes (DEGs) were obtained by obtaining the intersection points between these two datasets. IL-6, a common differential factor, was obtained by intersecting the DEGs with the General Olfactory Sensitivity Database (GOSdb) again. A model of AR mice with OD was developed by sensitizing with ovalbumin (OVA) to verify the reliability of IL-6 as a key factor of OD in AR and explore the potential mechanisms. Furthermore, a supernatant and microglia co-culture model of nasal mucosa epithelial cells stimulated by the allergen house dust mite extract Derp1 was established to identify the cellular and molecular mechanisms of IL-6-mediated OD in AR. RESULTS: The level of IL-6 in the nasal mucosa and olfactory bulb of AR mice with OD significantly increased and showed a positive correlation with the expression of olfactory bulb microglia marker Iba-1 and the severity of OD. In-vitro experiments showed that the level of IL-6 significantly increased in the supernatant after the nasal mucosa epithelial cells were stimulated by Derp1, along with significantly decreased barrier function of the nasal mucosa. The expression levels of neuroinflammatory markers IL-1ß and INOS increased after a conditioned culture of microglia with the supernatant including IL-6. Then knockdown (KD) of IL-6R by small interfering RNA (siRNA), the expression of IL-1ß and INOS significantly diminished. CONCLUSION: IL-6 plays a key role in the occurrence and development of OD in AR, which may be related to its effect on olfactory bulb microglia-mediated neuroinflammation.

Multimodal Action Quality Assessment.

Action quality assessment (AQA) is to assess how well an action is performed. Previous works perform modelling by only the use of visual information, ignoring audio information. We argue that although AQA is highly dependent on visual information, the audio is useful complementary information for improving the score regression accuracy, especially for sports with background music, such as figure skating and rhythmic gymnastics. To leverage multimodal information for AQA, i.e., RGB, optical flow and audio information, we propose a Progressive Adaptive Multimodal Fusion Network (PAMFN) that separately models modality-specific information and mixed-modality information. Our model consists of with three modality-specific branches that independently explore modality-specific information and a mixed-modality branch that progressively aggregates the modality-specific information from the modality-specific branches. To build the bridge between modality-specific branches and the mixed-modality branch, three novel modules are proposed. First, a Modality-specific Feature Decoder module is designed to selectively transfer modality-specific information to the mixed-modality branch. Second, when exploring the interaction between modality-specific information, we argue that using an invariant multimodal fusion policy may lead to suboptimal results, so as to take the potential diversity in different parts of an action into consideration. Therefore, an Adaptive Fusion Module is proposed to learn adaptive multimodal fusion policies in different parts of an action. This module consists of several FusionNets for exploring different multimodal fusion strategies and a PolicyNet for deciding which FusionNets are enabled. Third, a module called Cross-modal Feature Decoder is designed to transfer cross-modal features generated by Adaptive Fusion Module to the mixed-modality branch. Our extensive experiments validate the efficacy of the proposed method, and our method achieves state-of-the-art performance on two public datasets. Code is available at https://github.com/qinghuannn/PAMFN.

Incremental value of amyloid PET in a tertiary memory clinic setting in China.

INTRODUCTION: The objective of this study is to investigate the incremental value of amyloid positron emission tomography (Aß-PET) in a tertiary memory clinic setting in China. METHODS: A total of 1073 patients were offered Aß-PET using 18F-florbetapir. The neurologists determined a suspected etiology (Alzheimer's disease [AD] or non-AD) with a percentage estimate of their confidence and medication prescription both before and after receiving the Aß-PET results. RESULTS: After disclosure of the Aß-PET results, etiological diagnoses changed in 19.3% of patients, and diagnostic confidence increased from 69.3% to 85.6%. Amyloid PET results led to a change of treatment plan in 36.5% of patients. Compared to the late-onset group, the early-onset group had a more frequent change in diagnoses and a higher increase in diagnostic confidence. DISCUSSION: Aß-PET has significant impacts on the changes of diagnoses and management in Chinese population. Early-onset cases are more likely to benefit from Aß-PET than late-onset cases. HIGHLIGHTS: Amyloid PET contributes to diagnostic changes and its confidence in Chinese patients. Amyloid PET leads to a change of treatment plans in Chinese patients. Early-onset cases are more likely to benefit from amyloid PET than late-onset cases.

Plasma proteomic profiles predict future dementia in healthy adults.

The advent of proteomics offers an unprecedented opportunity to predict dementia onset. We examined this in data from 52,645 adults without dementia in the UK Biobank, with 1,417 incident cases and a follow-up time of 14.1 years. Of 1,463 plasma proteins, GFAP, NEFL, GDF15 and LTBP2 consistently associated most with incident all-cause dementia (ACD), Alzheimer's disease (AD) and vascular dementia (VaD), and ranked high in protein importance ordering. Combining GFAP (or GDF15) with demographics produced desirable predictions for ACD (area under the curve (AUC) = 0.891) and AD (AUC = 0.872) (or VaD (AUC = 0.912)). This was also true when predicting over 10-year ACD, AD and VaD. Individuals with higher GFAP levels were 2.32 times more likely to develop dementia. Notably, GFAP and LTBP2 were highly specific for dementia prediction. GFAP and NEFL began to change at least 10 years before dementia diagnosis. Our findings strongly highlight GFAP as an optimal biomarker for dementia prediction, even more than 10 years before the diagnosis, with implications for screening people at high risk for dementia and for early intervention.

How surface-to-volume ratio affects degradation of magnesium: in vitro and in vivo studies.

Despite the many studies carried out over the past decade to determine the biodegradation performance of magnesium and its alloys, few studies focused on the effect of altered surface area to volume ratio on in vitro and in vivo degradation rate and osteogenesis. Here, high purity magnesium cylindrical rods with gradient of surface area to volume ratio were processed by excavating different numbers of grooves on the side surface. The immersion test in SBF solution and the rat femoral condylar bone defect model were used to evaluate the degradation of magnesium rods in vitro and in vivo, respectively. We demonstrated that, the increased number of grooves on the HP magnesium surface represented a decrease in the percentage of residual volume over time, not necessarily an increase in absolute degradation volume or a regular change in corrosion rate. Furthermore, there were strong linear correlations between the relative degradation volume and the initial surface-to-volume ratio of HP magnesium rods both in vitro and in vivo. The difference in the slope of this relationship in vitro and in vivo might help to determine the possible range of in vivo degradation rates via in vitro data. In addition, the corrosion rate is more suitable for evaluating bone formation surrounding the different HP magnesium rods. Our findings in this work may facilitate adjusting the in vivo degradation and osteogenesis of different kinds of orthopedic implants made of the same magnesium-based material, and thus, accelerate the clinical popularization and application.

Community structure and species diversity dynamics of a subtropical evergreen broad-leaved forest in China: 2005 to 2020.

Here, we characterize the temporal and spatial dynamics of forest community structure and species diversity in a subtropical evergreen broad-leaved forest in China. We found that community structure in this forest changed over a 15-year period. Specifically, renewal and death of common species was large, with the renewal of individuals mainly concentrated within a few populations, especially those of Aidia canthioides and Cryptocarya concinna. The numbers of individual deaths for common species were concentrated in the small and mid-diameter level. The spatial distribution of community species diversity fluctuated in each monitoring period, showing a more dispersed diversity after the 15-year study period, and the coefficient of variation on quadrats increased. In 2010, the death and renewal of the community and the spatial variation of species diversity were different compared to other survey years. Extreme weather may have affected species regeneration and community stability in our subtropical monsoon evergreen broad-leaved forests. Our findings suggest that strengthening the monitoring and management of the forest community will help better understand the long- and short-term causes of dynamic fluctuations of community structure and species diversity, and reveal the factors that drive changes in community structure.

Young Plasma Attenuated Chronic Kidney Disease Progression after Acute Kidney Injury by Inhibiting Inflammation in Mice.

In the aged patients suffering from acute kidney injury, the risk for progression to chronic kidney disease and mortality is high. Aging accompanied by glomerulosclerosis, interstitial inflammation, and fibrosis might be one of the underlying mechanisms for vulnerability. In addition to sustained activation of the renin-angiotensin system, persistent chronic inflammation with tertiary lymphoid tissue formation is more common and is associated with disease progression in the aged kidney after acute injury. Based on recent laboratory evidence that young blood can rejuvenate the brain, muscle, and heart, we were intrigued by the possible protective effect of young plasma on acute kidney injury in aged mice. Here, we demonstrated that young plasma from 2-month-old mice could attenuate chronic kidney disease progression in 15-month-old mice subjected to acute kidney injury induced by ischemia-reperfusion. In the aged mice after acute kidney injury, young plasma administration decreased tubulointerstitial injury, fibrosis, and tertiary lymphoid tissue formation in kidneys assessed on day 28 after acute injury despite no significant beneficial effect on injury severity and survival. Mechanistically, young plasma inhibited angiotensin II-activated chemokines in pericytes that were responsible for tertiary lymphoid tissue formation. In summary, our data provide evidence that young plasma attenuates the transition from acute kidney injury to chronic kidney disease in aged mice. The therapeutic potential of young plasma infusion or exchange in the aged patients suffering acute kidney injury needs to be addressed in clinical trials.

Palmer Nodular Fasciitis Harboring a Novel SREBF1::USP6 Fusion Gene.

The diagnosis of low-grade fibroblastic/myofibroblastic tumors of acral sites can be challenging. These tumors encompass a diverse group of neoplasms with a spectrum of biologic potential ranges from benign to overtly malignant. They often demonstrate significant clinical, radiologic, and immunophenotypic overlap, in which the molecular phenotype may play an important diagnostic role to arrive at the final diagnosis. Herein, we report a case of soft tissue mass lesion presented on the palm of an adult patient for four months. Histologically, the tumor consisted of primarily low-grade spindle cells expressing smooth muscle actin. Molecular testing revealed a novel SREBF1::USP6 fusion gene, confirming the final diagnosis of nodular fasciitis and ultimately expanding its molecular profile. This case highlights the diagnostic value of single, cost-effective, targeted molecular panel to arrive at an accurate diagnosis and provide helpful therapeutic information.

Exome sequencing identifies genes associated with sleep-related traits.

Sleep is vital for human health and has a moderate heritability. Previous genome-wide association studies have limitations in capturing the role of rare genetic variants in sleep-related traits. Here we conducted a large-scale exome-wide association study of eight sleep-related traits (sleep duration, insomnia symptoms, chronotype, daytime sleepiness, daytime napping, ease of getting up in the morning, snoring and sleep apnoea) among 450,000 participants from UK Biobank. We identified 22 new genes associated with chronotype (ADGRL4, COL6A3, CLK4 and KRTAP3-3), daytime sleepiness (ST3GAL1 and ANKRD12), daytime napping (PLEKHM1, ANKRD12 and ZBTB21), snoring (WDR59) and sleep apnoea (13 genes). Notably, 20 of these genes were confirmed to be significantly associated with sleep disorders in the FinnGen cohort. Enrichment analysis revealed that these discovered genes were enriched in circadian rhythm and central nervous system neurons. Phenotypic association analysis showed that ANKRD12 was associated with cognition and inflammatory traits. Our results demonstrate the value of large-scale whole-exome analysis in understanding the genetic architecture of sleep-related traits and potential biological mechanisms.

Learning From Human Educational Wisdom: A Student-Centered Knowledge Distillation Method.

Existing studies on knowledge distillation typically focus on teacher-centered methods, in which the teacher network is trained according to its own standards before transferring the learned knowledge to a student one. However, due to differences in network structure between the teacher and the student, the knowledge learned by the former may not be desired by the latter. Inspired by human educational wisdom, this paper proposes a Student-Centered Distillation (SCD) method that enables the teacher network to adjust its knowledge transfer according to the student network's needs. We implemented SCD based on various human educational wisdom, e.g., the teacher network identified and learned the knowledge desired by the student network on the validation set, and then transferred it to the latter through the training set. To address the problems of current deficiency knowledge, hard sample learning and knowledge forgetting faced by a student network in the learning process, we introduce and improve Proportional-Integral-Derivative (PID) algorithms from automation fields to make them effective in identifying the current knowledge required by the student network. Furthermore, we propose a curriculum learning-based fuzzy strategy and apply it to the proposed PID control algorithm, such that the student network in SCD can actively pay attention to the learning of challenging samples after with certain knowledge. The overall performance of SCD is verified in multiple tasks by comparing it with state-of-the-art ones. Experimental results show that our student-centered distillation method outperforms existing teacher-centered ones.