Preoperative serum level of CA153 and a new model to predict the sub-optimal primary debulking surgery in patients with advanced epithelial ovarian cancer.

OBJECTIVE: The aim of this study was to establish a preoperative model to predict the outcome of primary debulking surgery (PDS) for advanced ovarian cancer (AOC) patients by combing Suidan predictive model with HE4, CA125, CA153 and ROMA index. METHODS: 76 AOC Patients in revised 2014 International Federation of Gynecology and Obstetrics (FIGO) stage III-IV who underwent PDS between 2017 and 2019 from Yunnan Cancer Hospital were included. Clinical data including the levels of preoperative serum HE4, CA125, CA153 and mid-lower abdominal CT-enhanced scan results were collected. The logistics regression analysis was performed to find factors associated with sub-optimal debulking surgery (SDS). The receiver operating characteristic curve was used to evaluate the predictive performances of selected variables in the outcome of primary debulking surgery. The predictive index value (PIV) model was constructed to predict the outcome of SDS. RESULTS: Optimal surgical cytoreduction was achieved in 61.84% (47/76) patients. The value for CA125, HE4, CA153, ROMA index and Suidan score was lower in optimal debulking surgery (ODS) group than SDS group. Based on the Youden index, which is widely used for evaluating the performance of predictive models, the best cutoff point for the preoperative serum HE4, CA125, CA153, ROMA index and Suidan score to distinguish SDS were 431.55 pmol/l, 2277 KU/L, 57.19 KU/L, 97.525% and 2.5, respectively. Patients with PIV≥5 may not be able to achieve optimal surgical cytoreduction. The diagnostic accuracy, NPV, PPV and specificity for diagnosing SDS were 73.7%, 82.9%, 62.9% and 72.3%, respectively. In the constructed model, the AUC of the SDS prediction was 0.770 (95% confidence interval: 0.654-0.887), P<0.001. CONCLUSION: Preoperative serum CA153 level is an important non-invasive predictor of primary SDS in advanced AOC, which has not been reported before. The constructed PIV model based on Suidan's predictive model plus HE4, CA125, CA153 and ROMA index can noninvasively predict SDS in AOC patients, the accuracy of this prediction model still needs to be validated in future studies.

BadLabel: A Robust Perspective on Evaluating and Enhancing Label-Noise Learning.

Label-noise learning (LNL) aims to increase the model's generalization given training data with noisy labels. To facilitate practical LNL algorithms, researchers have proposed different label noise types, ranging from class-conditional to instance-dependent noises. In this paper, we introduce a novel label noise type called BadLabel, which can significantly degrade the performance of existing LNL algorithms by a large margin. BadLabel is crafted based on the label-flipping attack against standard classification, where specific samples are selected and their labels are flipped to other labels so that the loss values of clean and noisy labels become indistinguishable. To address the challenge posed by BadLabel, we further propose a robust LNL method that perturbs the labels in an adversarial manner at each epoch to make the loss values of clean and noisy labels again distinguishable. Once we select a small set of (mostly) clean labeled data, we can apply the techniques of semi-supervised learning to train the model accurately. Empirically, our experimental results demonstrate that existing LNL algorithms are vulnerable to the newly introduced BadLabel noise type, while our proposed robust LNL method can effectively improve the generalization performance of the model under various types of label noise. The new dataset of noisy labels and the source codes of robust LNL algorithms are available at https://github.com/zjfheart/BadLabels.

Prediction of adverse events risk in patients with comorbid post-traumatic stress disorder and alcohol use disorder using electronic medical records by deep learning models.

BACKGROUND: Identifying co-occurring mental disorders and elevated risk is vital for optimization of healthcare processes. In this study, we will use DeepBiomarker2, an updated version of our deep learning model to predict the adverse events among patients with comorbid post-traumatic stress disorder (PTSD) and alcohol use disorder (AUD), a high-risk population. METHODS: We analyzed electronic medical records of 5565 patients from University of Pittsburgh Medical Center to predict adverse events (opioid use disorder, suicide related events, depression, and death) within 3 months at any encounter after the diagnosis of PTSD+AUD by using DeepBiomarker2. We integrated multimodal information including: lab tests, medications, co-morbidities, individual and neighborhood level social determinants of health (SDoH), psychotherapy and veteran data. RESULTS: DeepBiomarker2 achieved an area under the receiver operator curve (AUROC) of 0.94 on the prediction of adverse events among those PTSD+AUD patients. Medications such as vilazodone, dronabinol, tenofovir, suvorexant, modafinil, and lamivudine showed potential for risk reduction. SDoH parameters such as cognitive behavioral therapy and trauma focused psychotherapy lowered risk while active veteran status, income segregation, limited access to parks and greenery, low Gini index, limited English-speaking capacity, and younger patients increased risk. CONCLUSIONS: Our improved version of DeepBiomarker2 demonstrated its capability of predicting multiple adverse event risk with high accuracy and identifying potential risk and beneficial factors.

DeepBiomarker2: Prediction of Alcohol and Substance Use Disorder Risk in Post-Traumatic Stress Disorder Patients Using Electronic Medical Records and Multiple Social Determinants of Health.

Prediction of high-risk events amongst patients with mental disorders is critical for personalized interventions. We developed DeepBiomarker2 by leveraging deep learning and natural language processing to analyze lab tests, medication use, diagnosis, social determinants of health (SDoH) parameters, and psychotherapy for outcome prediction. To increase the model's interpretability, we further refined our contribution analysis to identify key features by scaling with a factor from a reference feature. We applied DeepBiomarker2 to analyze the EMR data of 38,807 patients from the University of Pittsburgh Medical Center diagnosed with post-traumatic stress disorder (PTSD) to determine their risk of developing alcohol and substance use disorder (ASUD). DeepBiomarker2 predicted whether a PTSD patient would have a diagnosis of ASUD within the following 3 months with an average c-statistic (receiver operating characteristic AUC) of 0.93 and average F1 score, precision, and recall of 0.880, 0.895, and 0.866 in the test sets, respectively. Our study found that the medications clindamycin, enalapril, penicillin, valacyclovir, Xarelto/rivaroxaban, moxifloxacin, and atropine and the SDoH parameters access to psychotherapy, living in zip codes with a high normalized vegetative index, Gini index, and low-income segregation may have potential to reduce the risk of ASUDs in PTSD. In conclusion, the integration of SDoH information, coupled with the refined feature contribution analysis, empowers DeepBiomarker2 to accurately predict ASUD risk. Moreover, the model can further identify potential indicators of increased risk along with medications with beneficial effects.

Enhancingß-Ga2O3-film ultraviolet detectors via RF magnetron sputtering with seed layer insertion onc-plane sapphire substrate.

Gallium oxide (Ga2O3) possesses a band gap of approximately 4.9 eV, aligning its detection wavelength within the solar-blind region, making it an ideal semiconductor material for solar-blind photodetectors. This study aims to enhance the performance of Ga2O3ultraviolet (UV) detectors by pre-depositing a Ga2O3seed layer on ac-plane sapphire substrate. The x-ray diffraction and x-ray photoelectron spectroscopy analyses validated that the deposited films, following high-temperature annealing, comprisedß-Ga2O3. Comparing samples with and without a 20 nm seed layer, it was found that the former exhibited fewer oxygen defects and substantially improved crystal quality. The incorporation of the seed layer led to the realization of detectors with remarkably low dark current (≤15.3 fA). Moreover, the photo-to-dark current ratio was enhanced by 30% (surpassing 1.3 × 104) and the response/recovery time reduced to 0.9 s/0.01 s, indicating faster performance. Furthermore, these detectors demonstrated higher responsivity (4.8 mA W-1), improved detectivity (2.49 × 1016Jones), and excellent solar-blind characteristics. This study serves as a foundational stepping toward achieving high-qualityß-Ga2O3thin film and UV detector arrays.

Prediction of Adverse Events Risk in Patients with Comorbid Post- Traumatic Stress Disorder and Alcohol Use Disorder Using Electronic Medical Records by Deep Learning Models.

Background: Prediction of high-risk events in mental disorder patients is crucial. In our previous study, we developed a deep learning model: DeepBiomarker by using electronic medical records (EMR) to predict suicide related event (SRE) risk in post-traumatic stress disorder (PTSD) patients. Methods: We applied DeepBiomarker2 through data integration of multimodal information: lab test, medication, co-morbidities, and social determinants of health. We analyzed EMRs of 5,565 patients from University of Pittsburgh Medical Center with a diagnosis of PTSD and alcohol use disorder (AUD) on risk of developing an adverse event (opioid use disorder, SREs, depression and death). Results: DeepBiomarker2 predicted whether a PTSD + AUD patient will have a diagnosis of any adverse events (SREs, opioid use disorder, depression, death) within 3 months with area under the receiver operator curve (AUROC) of 0.94. We found piroxicam, vilazodone, dronabinol, tenofovir, suvorexant, empagliflozin, famciclovir, veramyst, amantadine, sulfasalazine, and lamivudine to have potential to reduce risk. Conclusions: DeepBiomarker2 can predict multiple adverse event risk with high accuracy and identify potential risk and beneficial factors. Our results offer suggestions for personalized interventions in a variety of clinical and diverse populations.

Tumor suppressor p53 modulates activity-dependent synapse strengthening, autism-like behavior and hippocampus-dependent learning.

Synaptic potentiation underlies various forms of behavior and depends on modulation by multiple activity-dependent transcription factors to coordinate the expression of genes necessary for sustaining synaptic transmission. Our current study identified the tumor suppressor p53 as a novel transcription factor involved in this process. We first revealed that p53 could be elevated upon chemically induced long-term potentiation (cLTP) in cultured primary neurons. By knocking down p53 in neurons, we further showed that p53 is required for cLTP-induced elevation of surface GluA1 and GluA2 subunits of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR). Because LTP is one of the principal plasticity mechanisms underlying behaviors, we employed forebrain-specific knockdown of p53 to evaluate the role of p53 in behavior. Our results showed that, while knocking down p53 in mice does not alter locomotion or anxiety-like behavior, it significantly promotes repetitive behavior and reduces sociability in mice of both sexes. In addition, knocking down p53 also impairs hippocampal LTP and hippocampus-dependent learning and memory. Most importantly, these learning-associated defects are more pronounced in male mice than in female mice, suggesting a sex-specific role of p53 in these behaviors. Using RNA sequencing (RNAseq) to identify p53-associated genes in the hippocampus, we showed that knocking down p53 up- or down-regulates multiple genes with known functions in synaptic plasticity and neurodevelopment. Altogether, our study suggests p53 as an activity-dependent transcription factor that mediates the surface expression of AMPAR, permits hippocampal synaptic plasticity, represses autism-like behavior, and promotes hippocampus-dependent learning and memory.

Self-Assembled Complex Three-Phase Core-Shell Nanostructure of Au-CoFe2-TiN with a Magneto-Optical Coupling Effect.

Nanostructured plasmonic-magnetic metamaterials have gained great research interest due to their enhanced magneto-optical coupling effects. Here, we report a complex three-phase nanocomposite design combining ferromagnetic CoFe2 with plasmonic TiN and Au as a multifunctional hybrid metamaterial using either a cogrowth or a templated method. Via the first method of cogrowing three phases, three different morphologies of Au-CoFe2 core-shell nanopillars were formed in the TiN matrix. Via the second method of sequential deposition of a TiN-Au seed layer and a TiN-CoFe2 layer, highly ordered and uniform single-type core-shell nanopillars (i.e., the CoFe2 shell with a Au core) form in the TiN matrix. Both cogrowth and templated growth TiN-CoFe2-Au hybrid systems exhibit excellent epitaxial quality, hyperbolic dispersion, magnetic anisotropy, and a magneto-optical coupling effect. This study provides an effective approach for achieving highly uniform multiphase vertically aligned nanocomposite structures with well-integrated optical, magnetic, and coupling properties.

DeepBiomarker2: Prediction of alcohol and substance use disorder risk in post-traumatic stress disorder patients using electronic medical records and multiple social determinants of health.

Introduction: Prediction of high-risk events amongst patients with mental disorders is critical for personalized interventions. In our previous study, we developed a deep learning-based model, DeepBiomarker by utilizing electronic medical records (EMR) to predict the outcomes of patients with suicide-related events in post-traumatic stress disorder (PTSD) patients. Methods: We improved our deep learning model to develop DeepBiomarker2 through data integration of multimodal information: lab tests, medication use, diagnosis, and social determinants of health (SDoH) parameters (both individual and neighborhood level) from EMR data for outcome prediction. We further refined our contribution analysis for identifying key factors. We applied DeepBiomarker2 to analyze EMR data of 38,807 patients from University of Pittsburgh Medical Center diagnosed with PTSD to determine their risk of developing alcohol and substance use disorder (ASUD). Results: DeepBiomarker2 predicted whether a PTSD patient will have a diagnosis of ASUD within the following 3 months with a c-statistic (receiver operating characteristic AUC) of 0·93. We used contribution analysis technology to identify key lab tests, medication use and diagnosis for ASUD prediction. These identified factors imply that the regulation of the energy metabolism, blood circulation, inflammation, and microbiome is involved in shaping the pathophysiological pathways promoting ASUD risks in PTSD patients. Our study found protective medications such as oxybutynin, magnesium oxide, clindamycin, cetirizine, montelukast and venlafaxine all have a potential to reduce risk of ASUDs. Discussion: DeepBiomarker2 can predict ASUD risk with high accuracy and can further identify potential risk factors along with medications with beneficial effects. We believe that our approach will help in personalized interventions of PTSD for a variety of clinical scenarios.

The intergenerational succession and financialization of Chinese family enterprises: Considering the influence of heirs' growing experience.

As a mixed-methods research in economics and psychology, this study aimed to analyze the influence from the intergenerational succession on the financialization level including asset financialization and revenue financialization, and further test the moderating effect of the heirs' typical growing experience according to The Imprinting Theory, based on the 2009-2020 annual data of listed family enterprises of China. There were two key findings. First, the effect of Chinese family enterprises' intergenerational succession on asset financialization was positively significant while the effect on revenue financialization was not significant, indicating that the financialization behavior has not brought about effective financial profits. Second, among the heirs' typical growing experiences, their parents' entrepreneurial experience during their childhood, oversea study experience, and MBA education experience had the significantly positive moderating effects on the influence from intergeneration succession to asset financialization level of Chinese family enterprises, which was an important internal mechanism for the heirs to promote the financialization process of family enterprises.

Kernel Mixed Model for Transcriptome Association Study.

We introduce the python software package Kernel Mixed Model (KMM), which allows users to incorporate the network structure into transcriptome-wide association studies (TWASs). Our software is based on the association algorithm KMM, which is a method that enables the incorporation of the network structure as the kernels of the linear mixed model for TWAS. The implementation of the algorithm aims to offer users simple access to the algorithm through a one-line command. Furthermore, to improve the computing efficiency in case when the interaction network is sparse, we also provide the flexibility of computing with the sparse counterpart of the matrices offered in Python, which reduces both the computation operations and the memory required.

Numerical study on the wall-impinging diesel spray soot generation and oxidation in the cylinder under cold-start conditions of a diesel engine.

The combustion of wall-impinging diesel spray of heavy-duty diesel engines deteriorates combustion quality under cold-start conditions, making it difficult to control soot emissions. To investigate the causes of soot increase in the combustion of wall-impinging spray at low temperature and low speed starting conditions, the effect of the starting fuel mass on the soot formation and oxidation process was analyzed using a multidimensional computational fluid dynamics (CFD) model. The results show that the diesel spray is guided by the piston wall and the limited space, the spray impinged on the wall and the vapor-phase fuel flowed in the spray interaction zone. Thus, the soot mainly accumulates in the spray interaction zone, the region near the cylinder head and the bowl wall in the combustion chamber bowl. The soot from the vapor of deposited fuel film in the piston bowl wall and near wall region accumulates continuously in the after combustion stage, becoming the main source of soot emissions at the time of exhaust valve opening (EVO). Increasing the mass of starting fuel raises the mass of the rich mixture and wall-impinging fuel, which enhances the mismatch between fuel and air, resulting in higher soot generation, while soot is more difficult to be completely oxidized by OH radicals, and ultimately soot emissions increase significantly. It can be deduced that the engine-optimized injection strategy may mitigate the increase in soot emissions at high start-up fuel injection conditions.

Tunable physical properties in Bi-based layered supercell multiferroics embedded with Au nanoparticles.

Multiferroic materials are an interesting functional material family combining two ferroic orderings, e.g., ferroelectric and ferromagnetic orderings, or ferroelectric and antiferromagnetic orderings, and find various device applications, such as spintronics, multiferroic tunnel junctions, etc. Coupling multiferroic materials with plasmonic nanostructures offers great potential for optical-based switching in these devices. Here, we report a novel nanocomposite system consisting of layered Bi1.25AlMnO3.25 (BAMO) as a multiferroic matrix and well dispersed plasmonic Au nanoparticles (NPs) and demonstrate that the Au nanoparticle morphology and the nanocomposite properties can be effectively tuned. Specifically, the Au particle size can be tuned from 6.82 nm to 31.59 nm and the 6.82 nm one presents the optimum ferroelectric and ferromagnetic properties and plasmonic properties. Besides the room temperature multiferroic properties, the BAMO-Au nanocomposite system presents other unique functionalities including localized surface plasmon resonance (LSPR), hyperbolicity in the visible region, and magneto-optical coupling, which can all be effectively tailored through morphology tuning. This study demonstrates the feasibility of coupling single phase multiferroic oxides with plasmonic metals for complex nanocomposite designs towards optically switchable spintronics and other memory devices.

Effects of Different Pollination Methods on Oilseed Rape (Brassica napus) Plant Growth Traits and Rapeseed Yields.

Pollination success is essential for hybrid oilseed rape (OSR, Brassica napus) seed production, and the pollination method has some influences on the OSR plant growth traits. In order to explore the roles of different pollination methods, four pollination methods of "unmanned agricultural aerial system" (UAAS), "natural wind + UAAS" (NW+UAAS), "honeybee" (HB), and "no pollinators" (NP) were set in a hybrid OSR field to investigate their effects on OSR plant traits and rapeseed yields in this study. The control check (CK) area with natural wind (NW) pollination was set as a reference for comparison. The experiments were conducted continuously for 20 days during the OSR plant early to full-bloom stage. The results based on the evaluated OSR plants showed that the growth traits and the rapeseed yields exhibited some differences under different pollination methods. The average plant height under NP pollination was maximum, which was 231.52 cm, while the average plant heights under the other pollination methods exhibited nearly no difference. Except for the HB pollination, the average first-branch heights of the evaluated plants all exceeded 100 cm under the other pollination methods. The average once branch quantity of all the evaluated plants under different pollination methods was 5-7. The average number of effective siliques per plant varied greatly. The average quantity of effective siliques in each OSR plant was about 160 under UAAS, NW+UAAS, and NW pollination, about 100 under HB pollination, and only 2.12 under NP pollination. The thousand-rapeseed weight was 7.32 g under HB pollination, which was the highest of all of the pollination areas. In terms of rapeseed yield, the average rapeseed yields per plant were all more than 10 g, except for the one under NP pollination; the yield per hectare was highest under NW+UAAS pollination, reaching 4741.28 kg, and the yield under NP pollination was lowest, which was only 360.39 kg. The research results provide technical support for supplementary pollination in hybrid OSR seed production.

Intrinsic ferroelectricity in Y-doped HfO2 thin films.

Ferroelectric HfO2-based materials hold great potential for the widespread integration of ferroelectricity into modern electronics due to their compatibility with existing Si technology. Earlier work indicated that a nanometre grain size was crucial for the stabilization of the ferroelectric phase. This constraint, associated with a high density of structural defects, obscures an insight into the intrinsic ferroelectricity of HfO2-based materials. Here we demonstrate that stable and enhanced polarization can be achieved in epitaxial HfO2 films with a high degree of structural order (crystallinity). An out-of-plane polarization value of 50 µC cm-2 has been observed at room temperature in Y-doped HfO2(111) epitaxial thin films, with an estimated full value of intrinsic polarization of 64 µC cm-2, which is in close agreement with density functional theory calculations. The crystal structure of films reveals the Pca21 orthorhombic phase with small rhombohedral distortion, underlining the role of the structural constraint in stabilizing the ferroelectric phase. Our results suggest that it could be possible to exploit the intrinsic ferroelectricity of HfO2-based materials, optimizing their performance in device applications.

DeepBiomarker: Identifying Important Lab Tests from Electronic Medical Records for the Prediction of Suicide-Related Events among PTSD Patients.

Identifying patients with high risk of suicide is critical for suicide prevention. We examined lab tests together with medication use and diagnosis from electronic medical records (EMR) data for prediction of suicide-related events (SREs; suicidal ideations, attempts and deaths) in post-traumatic stress disorder (PTSD) patients, a population with a high risk of suicide. We developed DeepBiomarker, a deep-learning model through augmenting the data, including lab tests, and integrating contribution analysis for key factor identification. We applied DeepBiomarker to analyze EMR data of 38,807 PTSD patients from the University of Pittsburgh Medical Center. Our model predicted whether a patient would have an SRE within the following 3 months with an area under curve score of 0.930. Through contribution analysis, we identified important lab tests for suicide prediction. These identified factors imply that the regulation of the immune system, respiratory system, cardiovascular system, and gut microbiome were involved in shaping the pathophysiological pathways promoting depression and suicidal risks in PTSD patients. Our results showed that abnormal lab tests combined with medication use and diagnosis could facilitate predicting SRE risk. Moreover, this may imply beneficial effects for suicide prevention by treating comorbidities associated with these biomarkers.

Trade-offs of Linear Mixed Models in Genome-Wide Association Studies.

Motivated by empirical arguments that are well known from the genome-wide association studies (GWAS) literature, we study the statistical properties of linear mixed models (LMMs) applied to GWAS. First, we study the sensitivity of LMMs to the inclusion of a candidate single nucleotide polymorphism (SNP) in the kinship matrix, which is often done in practice to speed up computations. Our results shed light on the size of the error incurred by including a candidate SNP, providing a justification to this technique to trade off velocity against veracity. Second, we investigate how mixed models can correct confounders in GWAS, which is widely accepted as an advantage of LMMs over traditional methods. We consider two sources of confounding factors-population stratification and environmental confounding factors-and study how different methods that are commonly used in practice trade off these two confounding factors differently.

Strong Interfacial Coupling of Tunable Ni-NiO Nanocomposite Thin Films Formed by Self-Decomposition.

The next-generation spintronic devices including memristors, tunneling devices, or stochastic switching exert surging demands on magnetic nanostructures with novel coupling schemes. Taking advantage of a phase decomposition mechanism, a unique Ni-NiO nanocomposite has been demonstrated using a conventional pulsed laser deposition technique. Ni nanodomains are segregated from NiO and exhibit as faceted "emerald-cut" morphologies with tunable dimensions affected by the growth temperature. The sharp interfacial transition between ferromagnetic (002) Ni and antiferromagnetic (002) NiO, as characterized by high-resolution transmission electron microscopy, introduces a strong exchange bias effect and magneto-optical coupling at room temperature. In situ heating-cooling X-ray diffraction (XRD) study confirms an irreversible phase transformation between Ni and NiO under ambient atmosphere. Synthesizing highly functional two-phase nanocomposites with a simple bottom-up self-assembly via such a phase decomposition mechanism presents advantages in terms of epitaxial quality, surface coverage, interfacial coupling, and tunable nanomagnetism, which are valuable for new spintronic device implementation.

The Effect of Product Image Dynamism on Purchase Intention for Online Aquatic Product Shopping: An EEG Study.

PURPOSE: The normalization of epidemic prevention and control triggered a fierce scuffle in the e-commerce of fresh food, as well as for aquatic products online shopping. The main difficulty for consumers to buy fresh food online has always been information asymmetry. Previous study reported that the image is still the primary information source to address information asymmetry. Yet, few studies have focused on the image presentation of aquatic products in e-commerce. The current study aims to probe the effect of perceived movement of e-commerce pictures on purchase intention of aquatic products. Further, we examine how consumers' cognitive conflict and emotion occur when purchasing specific aquatic products with different image dynamism. METHODS: Twenty-eight subjects participated in an experiment with a 2-level product category (fresh vs frozen) × 2-level image dynamism (static vs dynamic) design. During the experiment, participants were asked to rate their purchase intention after they browse the experimental stimulus. We recorded subjects' electroencephalograms (EEGs) throughout the experiment. RESULTS: Behaviorally, participants' purchase intention for the dynamic image was significantly greater than that for the static image, regardless of aquatic product categories. At the neural level, we found that dynamic image produced less cognitive conflict and aroused consumers' positive feelings, which were reflected in decreased N2 amplitudes and latency as well as increased LPP (late positive potential) amplitude, respectively. This effect was enhanced for fresh aquatic products. CONCLUSION: Although picture dynamism only increases perceived movement, it can still induce positive emotions toward the product and lead to a greater purchase intention. The current study emphasized the value of the neuroscience method in revealing consumer cognition and emotion duration product evaluation.

Active learning to classify macromolecular structures in situ for less supervision in cryo-electron tomography.

MOTIVATION: Cryo-Electron Tomography (cryo-ET) is a 3D bioimaging tool that visualizes the structural and spatial organization of macromolecules at a near-native state in single cells, which has broad applications in life science. However, the systematic structural recognition and recovery of macromolecules captured by cryo-ET are difficult due to high structural complexity and imaging limits. Deep learning-based subtomogram classification has played critical roles for such tasks. As supervised approaches, however, their performance relies on sufficient and laborious annotation on a large training dataset. RESULTS: To alleviate this major labeling burden, we proposed a Hybrid Active Learning (HAL) framework for querying subtomograms for labeling from a large unlabeled subtomogram pool. Firstly, HAL adopts uncertainty sampling to select the subtomograms that have the most uncertain predictions. This strategy enforces the model to be aware of the inductive bias during classification and subtomogram selection, which satisfies the discriminativeness principle in AL literature. Moreover, to mitigate the sampling bias caused by such strategy, a discriminator is introduced to judge if a certain subtomogram is labeled or unlabeled and subsequently the model queries the subtomogram that have higher probabilities to be unlabeled. Such query strategy encourages to match the data distribution between the labeled and unlabeled subtomogram samples, which essentially encodes the representativeness criterion into the subtomogram selection process. Additionally, HAL introduces a subset sampling strategy to improve the diversity of the query set, so that the information overlap is decreased between the queried batches and the algorithmic efficiency is improved. Our experiments on subtomogram classification tasks using both simulated and real data demonstrate that we can achieve comparable testing performance (on average only 3% accuracy drop) by using less than 30% of the labeled subtomograms, which shows a very promising result for subtomogram classification task with limited labeling resources. AVAILABILITY AND IMPLEMENTATION: https://github.com/xulabs/aitom. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.