Search

Several widely recognized metabolites play a role in regulating the pathophysiological processes of various disorders. Nonetheless, the lack of effective biomarkers for the early diagnosis of polycystic ovarian syndrome (PCOS) and premature ovarian failure (POF) has led to the discovery of serum-based metabolic biomarkers for these disorders. We aimed to identify various differentially expressed metabolites (DEMs) through serum-based metabolic profiling in patients with PCOS and POF and in healthy individuals by using liquid chromatography–mass spectrometry analysis. Furthermore, heatmap clustering, correlation, and Z-score analyses were performed to identify the top DEMs. Kyoto Encyclopedia of Genes and Genomes enriched pathways of DEMs were determined using metabolite-based databases. Moreover, the clinical significance of these DEMs was evaluated on the basis of area under the receiver operating characteristic curve. Significantly dysregulated expressions of several metabolites were observed in the intergroup comparisons of the PCOS, POF, and healthy control groups. Furthermore, 6 DEMs were most frequently observed among the three groups. The expressions of these DEMs were not only directly correlated but also exhibited potential significance in patients with PCOS and POF. Novel metabolites with up/downregulated expressions can be discovered in patients with PCOS and POF using serum-based metabolomics; these metabolites show good diagnostic performance and can act as effective biomarkers for the early detection of PCOS and POF. Furthermore, these metabolites might be involved in the pathophysiological mechanisms of PCOS and POF via interplay with corresponding genes.

Aspect level sentiment classification aims to extract fine-grained sentiment expressed towards specific aspects from a sentence. The key to this task lies in connecting aspects and their respective sentiment contexts. Existing methods measure the dependency weights between aspects and context words via either the semantic similarity between words captured by attention mechanism or the structural proximity between words in syntactic structures. However, methods in both groups fail to fully exploit explicit syntactic dependency, which we argue should be critical to identify sentiment contexts. In this paper, we propose a novel syntactic-dependency-based attention network (SDATT) to incorporate explicit syntactic dependency for aspect level sentiment classification. SDATT first models the dependency path between each word and the aspect to characterize aspect-oriented syntactic representation of each word. The generated syntactic representations are later fed into the attention layer to help infer the dependency weights for sentiment prediction. Experimental results on five benchmark datasets show the superior performance of the proposed model over state-of-the-art baselines.

Small molecule organic photovoltaics (SMPVs) were prepared by utilizing liquid crystalline donor material BTR-Cl and two similar optical bandgap non-fullerene acceptor materials BTP-BO-4F and Y6. The BTP-BO-4F and Y6 have the similar optical bandgap and different absorption coefficients. The corresponding binary SMPVs exhibit different short circuit current density (JSC) (20.38 vs. 23.24 mA cm−2), and fill factor (FF) (70.77% vs. 67.21%). A 14.46% power conversion efficiency (PCE) is acquired in ternary SMPVs with 30 wt% Y6, companied with a JSC of 24.17 mA cm−2, a FF of 68.78% and an open circuit voltage (VOC) of 0.87 V. The improvement on PCE of ternary SMPVs should originate from the well trade-off between phase separation and photon harvesting of ternary active layers by incorporating 30 wt% Y6 in acceptors. This work may deliver insight onto the improved performance of SMPVs by superposing the superiorities of binary SMPVs with similar optical bandgap acceptors into one ternary cell.

According to the model of sliding ferroelectricity we proposed in 2017, a type of vertical polarization switchable via interlayer sliding may exist in a series of two-dimensional van der Waals bilayers and multilayers. Such intriguing ferroelectricity has been recently experimentally confirmed in BN and WTe2 bilayer/multilayer systems. However, they are respectively insulators and metals with weak polarizations. Here, we search for a combination of high-mobility semiconductors and ferroelectricity with relatively high polarizations, and we note that MoSi2N4 monolayer has been recently fabricated to centimeter scale based on MoN2 monolayer (Science 2020, 369, 670). We show first-principles evidence of strong interlayer sliding ferroelectricity in high-mobility semiconducting MoA2N4 (A = Si or Ge) bilayers and multilayers. They possess tunable bandgaps within the desirable range for nanoelectronics, with the current highest polarizations in sliding ferroelectrics known to date due to strong interlayer van der Waals interactions. Such a successful combination of these properties may render the long-sought and efficient computing-in-memory possible. Their high polarizations induce strong Moire potential and unique band alignments for exciton trapping in twisted MoA2N4 bilayer. The varying charge distribution of different stacking in MoN2/MoSi2N4 hetero-bilayer also gives rise to a high alternating voltage when used as a nanogenerator, suggesting promising potential for energy harvesting.

Over 200 nm thickness of active layers are essential for achieving efficient organic photovolatics (OPVs) using the roll-to-roll (R2R) large-scale production technology. In this study, a series of OPVs with 300 nm thick active layers were prepared with PM6:BTP-4F-12:BP-4F as active layers. The PM6:BTP-4F-12- and PM6:BP-4F-based binary OPVs achieve power conversion efficiencies (PCE) of 14.62% and 14.14%, respectively, with 300 nm thick active layers, also exhibiting complementary photovoltaic parameters. A PCE of 15.63% is achieved by the optimized ternary thick-film OPVs by combining the superiorities of two binary thick-film OPVs, deriving from a short circuit current density (JSC) of 26.57 mA cm−2, an open circuit voltage (VOC) of 0.840 V and a fill factor (FF) of 70.03%. The performance improvement of ternary thick-film OPVs should be mainly ascribed to the optimized molecular arrangement and phase separation for efficient charge transport in ternary thick active layers, which can be confirmed from the significantly improved FFs of ternary thick-film OPVs. This study indicates that the ternary strategy has great potential in preparing efficient thick-film OPVs.

Ternary organic photovoltaics (OPVs) were constructed with one wide-band-gap donor PM6 and two A-D-A-type acceptors (M-series M36 and MQ5) with similar chemical structures. Power conversion efficiency (PCE) of the optimal ternary OPVs reaches 17.24% with 20 wt % MQ5 content, arising from a simultaneously increased short circuit current density (JSC) of 25.36 mA cm-2 and a fill factor (FF) of 76.02% as compared to those of two binary OPVs. The photon harvesting of ternary active layers can be maximized by adjusting the MQ5 content by reason of the complementary absorption spectra of M36 and MQ5. The molecular arrangement of PM6 and M36 can be collectively optimized by introducing an appropriate amount of MQ5 as a morphology regulator for facilitating effective charge transportation in ternary active layers. The improved photon harvesting and charge transport in active layers should be two important factors responsible for JSC and FF improvement of optimal ternary OPVs, respectively. More than an 8.8% improvement of PCE is achieved in ternary OPVs with an appropriate amount of MQ5 as the photon-harvesting enhancer and morphology regulator. The huge potential of A-D-A-type materials in constructing highly efficient OPVs can be further exploited based on a ternary strategy.

The ternary strategy exhibits great potential in optimizing the photon harvesting and phase separation of active layers. In this work, non-fullerene MF1 was selected as the third component to prepare efficient ternary organic solar cells (OSCs) by finely optimizing the MF1 content in the acceptors. The optimized power conversion efficiency (PCE) of 15.31% is achieved in the ternary OSCs with 20 wt% MF1 content in the acceptors and 100 nm active layer thickness, also exhibiting a relatively high fill factor (FF) of 78.05%. The relatively high FF indicates efficient charge transport and collection in the optimized ternary OSCs, which should be beneficial to achieve efficient thick-film OSCs. It is highlighted that a PCE of 14.57% is achieved in the optimized ternary OSCs with 300 nm thick active layers compatible with the roll-to-roll (R2R) large-scale printing process. To date, high performance thick-film ternary non-fullerene OSCs have seldom been reported. This work indicates that the thick-film ternary strategy has great potential in achieving efficient large-scale OSCs.

Due to their unique merits, asymmetrical fused-ring electron acceptors (FREAs) have received increasing attention in the field of organic photovoltaics (OPVs). Herein, we designed and synthesized an asymmetrical FREA, namely AY6, by employing a cross-conceptual strategy which combines the structural features of asymmetrical IT6-4F with high-performance Y6. Compared to Y-shape Y6, X-shape AY6 shows poor crystallization properties and thus a significantly blue-shifted absorption. However, it is precisely because of this unfavorable self-stacking effect that AY6 has an opportunity to pack with PM6 and form a new intermixed phase. As a result, the PM6:AY6 blend film exhibits a perfect face-on orientation with stronger π–π stacking than the PM6:Y6 blend film, which facilitates efficient exciton dissociation and balanced carrier transport, and therefore higher FF. In general, such a cross-conceptual design strategy is an effective method which not only boosts the efficiency of asymmetrical FREA-based OSCs but also increases the fill factors of OSCs based on Y6-like FREAs.

A series of ternary organic photovoltaics (OPVs) are fabricated with one wide bandgap polymer D18-Cl as donor, and well compatible Y6 and Y6-1O as acceptor. The open-circuit-voltage (VOC) of ternary OPVs is monotonously increased along with the incorporation of Y6-1O, indicating that the alloy state should be formed between Y6 and Y6-1O due to their excellent compatibility. The energy loss can be minimized by incorporating Y6-1O, leading to the VOC improvement of ternary OPVs. By finely adjusting the Y6-1O content, a power conversion efficiency of 17.91% is achieved in the optimal ternary OPVs with 30 wt% Y6-1O in acceptors, resulting from synchronously improved short-circuit-current density (JSC) of 25.87 mA cm−2, fill factor (FF) of 76.92% and VOC of 0.900 V in comparison with those of D18-Cl : Y6 binary OPVs. The JSC and FF improvement of ternary OPVs should be ascribed to comprehensively optimal photon harvesting, exciton dissociation and charge transport in ternary active layers. The more efficient charge separation and transport process in ternary active layers can be confirmed by the magneto-photocurrent and impedance spectroscopy experimental results, respectively. This work provides new insight into constructing highly efficient ternary OPVs with well compatible Y6 and its derivative as acceptor., By employing the well compatible Y6 : Y6-1O as acceptor and wide bandgap polymer D18-Cl as donor, the PCE of 17.91% is achieved in optimized ternary OPVs, resulting from the simultaneously improved JSC, FF and VOC.

Graph neural networks (GNNs) achieve remarkable success in graph-based semi-supervised node classification, leveraging the information from neighboring nodes to improve the representation learning of target node. The success of GNNs at node classification depends on the assumption that connected nodes tend to have the same label. However, such an assumption does not always work, limiting the performance of GNNs at node classification. In this paper, we propose label-consistency based graph neural network (LC-GNN), leveraging node pairs unconnected but with the same labels to enlarge the receptive field of nodes in GNNs. Experiments on benchmark datasets demonstrate the proposed LC-GNN outperforms traditional GNNs in graph-based semi-supervised node classification. We further show the superiority of LC-GNN in sparse scenarios with only a handful of labeled nodes.

The power conversion efficiency (PCE) of 13.00% was achieved in PBDB-T:Y16-based polymer solar cells (PSCs). On this basis, PCE of ternary PSCs was improved to 14.11% by incorporating 15 wt% MeIC1 in acceptors, resulting from simultaneously enhanced short circuit current (JSC) of 22.76 mA cm−2 and fill factor (FF) of 68.22%. The observed 14.11% PCE is among the highest values for all ternary PSCs. Y16 and MeIC1 preferred to form an alloyed acceptor due to good compatibility, which is beneficial to the formation of efficient electron transport channels in ternary active layers. Photon harvesting, exciton dissociation and charge transport could be synergistically optimized by incorporating 15 wt% MeIC1 in the acceptors. The optical field and photogenerated exciton distribution in active layers were calculated according to their intrinsic properties, which could provide more intuitive evidence of JSC and FF improvement. The photogenerated exciton distribution in active layers could also be optimized by employing a ternary strategy, which was beneficial for better balance charge collection efficiency and for achieving high FF of the ternary PSCs. This work further demonstrates that ternary strategies have great potential for improving the PSC performance by simultaneously optimizing photon harvesting and photogenerated exciton distribution in active layers.

Social network platforms have gradually transformed into social media in recent years, which greatly eases information diffusion. Meanwhile, it also makes the task of popularity prediction for online information more challenging. Traditional popularity prediction methods include supervised methods based on feature engineering and popularity dynamics modeling methods based on stochastic processes. The latter group of methods has drawn a wide attention as it works better for predicting individual items popularity. However, the existing methods have ignored the decentralization characteristic of diffusion cascades. In this paper, we investigate the decentralized structure of diffusion cascades, based on the Weibo dataset, and propose to model cascade dynamics by mixing a bunch of reinforced Poisson processes (RPP). The overall diffusion cascade is a mixture of several sub-cascades, each of which can be modeled by RPP. Experimental results on real world datasets demonstrate our proposed models superiority in both characterizing popularity dynamics and predicting future popularity.

Semitransparent solar cells not only can generate electricity but also exhibit promising application in facades and roofs of building, as well as the coatings of vehicle. Ideal semitransparent solar cells should have a high utilization efficiency of ultraviolet and near infrared photons, and maintain an appropriate balance between absorption and transparency in visible light range. Organic semiconducting materials with adjustable optical band gap own great potential in preparing highly efficient semitransparent organic solar cells (OSCs). There are great challenges in achieving semitransparent electrodes with good conductivity and high transmittance in visible light range, as well as high reflectance in ultraviolet and near infrared range. In this critical review, we summarize the recent progress of semitransparent OSCs, concentrating on organic semiconducting materials, semitransparent top electrode and device engineering. Semitransparent OSCs have attracted more and more attention due to its great potential in energy source in door, electricity-generating windows, building-integrated photovoltaics and agricultural greenhouse.

Solution processed interfacial layers are commonly employed in bulk heterojunction organic solar cells (OSCs) for better charge collection. PDIN interfacial layers were prepared by employing a static or dynamic spin coating method from PDIN methanol solution, and defined as the S-PDIN or D-PDIN layer. The OSCs with a S-PDIN layer exhibit 13.88% power conversion efficiency (PCE) with a virtual high short circuit density (JSC) of 26.45 mA cm-2 and relatively low fill factor (FF) of 58.94% during the current density versus voltage (J-V) measurement without a shadow mask. 12.56% PCE is achieved for OSCs with a D-PDIN layer, along with a JSC of 18.85 mA cm-2 and FF of 74.88%. Over 77% FFs are obtained for OSCs with a S-PDIN or D-PDIN layer during J-V measurement with a shadow mask, and both OSCs exhibit a very similar JSC and PCE. The virtual high JSCs and relatively low FF of OSCs with a S-PDIN layer may be due to the enhanced conductivity of PEDOT:PSS during preparation of the PDIN layer by the SSC method, which can be further confirmed from the OSCs with a methanol treated PEDOT:PSS layer. This work indicates that a well-balanced JSC and FF should be an important evaluating indicator for efficient OSCs, and an appropriate shadow mask is necessary to measure the J-V curves of OSCs with a solution processed interfacial layer.

Predicting the popularity of online content on social platforms is an important task for both researchers and practitioners. Previous methods mainly leverage demographics, temporal and structural patterns of early adopters for popularity prediction. However, most existing methods are less effective to precisely capture the cascading effect in information diffusion, in which early adopters try to activate potential users along the underlying network. In this paper, we consider the problem of network-aware popularity prediction, leveraging both early adopters and social networks for popularity prediction. We propose to capture the cascading effect explicitly, modeling the activation state of a target user given the activation state and influence of his/her neighbors. To achieve this goal, we propose a novel method, namely CoupledGNN, which uses two coupled graph neural networks to capture the interplay between node activation states and the spread of influence. By stacking graph neural network layers, our proposed method naturally captures the cascading effect along the network in a successive manner. Experiments conducted on both synthetic and real-world Sina Weibo datasets demonstrate that our method significantly outperforms the state-of-the-art methods for popularity prediction.

The Bitcoin market becomes the focus of the economic market since its birth, and it has attracted wide attention from both academia and industry. Due to the absence of regulations in the Bitcoin market, it may be easier to bring some kinds of illegal behaviors. Thus, it raises an interesting question: Is there abnormity or illegal behavior in Bitcoin platforms? To answer this question, we investigate the abnormity in five leading Bitcoin platforms. By analyzing the financial index, i.e. the normalized logarithmic price return, we find that the properties of price return in bitFlyer are completely different from others. To find the possible reasons, we find that the abnormal ask price and bid price appear simultaneously in bitFlyer, which may be potentially linked to either price manipulation or money laundering. It verifies our conjecture that there may be abnormity or price manipulation in Bitcoin platforms. Furthermore, our findings in price return could also provide an innovative and effective method to detect the abnormity in Bitcoin platforms.

Background: Recurrence is a major cause of cancer-related deaths in colorectal cancer (CRC) patients, but the current strategies are limited to predict this clinical behavior. Our aim is to develop a recurrence prediction model based on long non-coding RNAs (lncRNAs) in exosomes of serum to improve the prediction accuracy. Methods: High-throughput lncRNAs microarray and reverse transcription quantitative real-time PCR were used to identify the recurrence-associated lncRNAs in 30 matched CRC tissues and sera. The recurrence prediction model was built with multivariable Cox analysis in 150-patient training set, and evaluated in an independent 203-patient test set using ROC, Kaplan-Meier, and COX analysis. Findings: In discovery phase, 11 lncRNAs were found to be associated with CRC recurrence in tissues, and 9 of them were correlated with their expression levels of serum exosomes. In training phase, a model based on 5-exosomal lncRNAs (exolncRNAs) panel was constructed, and showed high distinguish capability for recurrent CRC patients. ROC showed the panel was superior to serum CEA and CA19-9 in prediction of CRC recurrence. In both training and test sets, high-risk patients defined by the 5-exolncRNAs panel had poor recurrence free and overall survival. And, it was an independent factor for CRC prognosis. Interpretation: The 5-exolncRNAs panel robustly stratifies CRC patients' risk of recurrence, enabling more accurate prediction of prognosis. Funding Statement: Jinan Science and Technology Development Plan (201805003, 201805061), Shandong Medical and Health Technology Development Project (2018WSB20002), Shandong Key Research and Development Program (2018GSF118104), and National Natural Science Foundation of China (81301506). Declaration of Interests: The authors declared: "There are no relevant conflicts of interest." Ethics Approval Statement: This study was approved by the Ethics Committee of Shandong Provincial Third Hospital and Qilu Hospital of Shandong University, and written informed consent was obtained from each patient.

Ternary strategy has been confirmed as an efficient method to improve the power conversion efficiency (PCE) of organic photovoltaics (OPVs). The 15.7% PCE is achieved from PM6:Y6 based binary OPVs. One nonfullerene acceptor Br-ITIC and fullerene derivative PC71BM are selected as the third component on the basis of efficient binary OPVs, respectively. The optimized ternary OPVs exhibit 16.4% and 16.2% PCE with Br-ITIC and PC71BM as the third component, respectively, corresponding to the short circuit current density (JSC) of 25.5 mA cm−2 vs. 25.6 mA cm−2, open circuit voltage (VOC) of 0.854 V vs. 0.836 V and fill factor (FF) of 75.1% vs. 75.6%. The advantage on photovoltaic parameters of two ternary OPVs may be recombined into one cell by employing PC71BM as the fourth component. A 16.8% PCE is achieved from the optimized quaternary OPVs, resulting from the further increased JSC of 25.8 mA cm−2 and FF of 76.4% compared with the optimized ternary OPVs. The third party certificated PCE of quaternary OPVs is 16.2%. In comparison to 15.7% PCE of the binary OPVs, about 4.5% and 7.0% PCE improvement are step-by-step achieved from the optimized ternary and quaternary OPVs, respectively. Multi-components strategy may provide enough room to achieve highly efficient OPVs.

An ability of modeling and predicting the cascades of resharing is crucial to understanding information propagation and to launching campaign of viral marketing. Conventional methods for cascade prediction heavily depend on the hypothesis of diffusion models, e.g., independent cascade model and linear threshold model. Recently, researchers attempt to circumvent the problem of cascade prediction using sequential models (e.g., recurrent neural network, namely RNN) that do not require knowing the underlying diffusion model. Existing sequential models employ a chain structure to capture the memory effect. However, for cascade prediction, each cascade generally corresponds to a diffusion tree, causing cross-dependence in cascade---one sharing behavior could be triggered by its non-immediate predecessor in the memory chain. In this paper, we propose to an attention-based RNN to capture the cross-dependence in cascade. Furthermore, we introduce a \emph{coverage} strategy to combat the misallocation of attention caused by the memoryless of traditional attention mechanism. Extensive experiments on both synthetic and real world datasets demonstrate the proposed models outperform state-of-the-art models at both cascade prediction and inferring diffusion tree.

Predicting the popularity of online content is highly valuable in many applications and has been studied for several years. However, existing models either work in population level---all messages are assumed to follow similar popularity dynamics, lacking flexibility to capture the intrinsic complexity of popularity dynamics, or work in individual level---the popularity dynamics of messages are independent of each other, failing to leverage other messages to improve prediction accuracy. In this paper, we propose a divide and conquer framework for popularity prediction. We first divide messages into groups, anticipating each group of messages follow similar popularity dynamics, and then, we train a group-specific model for the messages of each group. Experiments demonstrate that group-specific models improve the population-level models by about 30% and outperform state-of-the-art individual-level model.

Modeling and predicting retweeting dynamics in social media has important implications to an array of applications. Existing models either fail to model the triggering effect of retweeting dynamics, e.g., the model based on reinforced Poisson process, or are hard to be trained using only the retweeting dynamics of individual tweet, e.g., the model based on self-exciting Hawkes process. In this paper, motivated by the observation that each retweeting dynamics is generally dominated by a handful of key nodes that separately trigger a high number of retweets, we propose a mixture process to model and predict retweeting dynamics, with each subprocess capturing the retweeting dynamics initiated by a key node. Experiments demonstrate that the proposed model outperforms the state-of-the-art model.

Since human face is often not clear in the surveillance video, this paper proposes a retrieval method on the whole body with fine-grained feature extraction. This method first extracts foreground region of human movement based on Gaussian mixed model (GMM). The human body is divided into two parts, head and below the head, based on the human morphological features and skin color. There are three parts coupled with the whole body, and then, we extract color feature for each part. Secondly, the human body is divided into front and back feature samples according to the a priori knowledge. Calculate the gap between the retrieval eigenvectors and the eigenvectors of the target body, then determine whether match. The experimental results show that this method maintains better retrieval precision when the recall rate is high. This paper target retrieve is real time on video.

The feature extraction is the most critical step in image retrieval. Among various local feature extraction methods, scale-invariant feature transform (SIFT) has been proven to be the most robust local invariant feature descriptor, which is widely used in the field of image matching and retrieval. However, the SIFT algorithm has a disadvantage that the algorithm will produce a large number of feature points and is not suited for widely using in the field of image retrieval. Firstly, a novel significant measure algorithm is proposed in this paper, and the regions of interest in images are obtained. Then, SIFT features are extracted from salient regions, reducing the number of SIFT features. Our algorithm also abstracts color features from salient regions, and this method overcomes SIFT algorithm’s drawback that could not reflect image’s color information. The experiments demonstrate that the integrated visual saliency analysis-based feature selection algorithm provides significant benefits both in retrieval accuracy and in speed.

Microblog has gained more and more users around the world, the popularity of which makes information spreading in microblog the most important and influential activities on the Internet. Therefore, search in microblog is of the most significant issue for both academic and industrial world. Search in webpages has been studied for several decades, but as for microblog it is still an open and brand new question for everyone. Search in microblog is more difficult than that in traditional webpages because of the sparseness of the messages. Search functions in current microblogging services simply match microblog messages with query words, which cannot guarantee the correlation between the retrieving messages and the users’ intention. We introduce the concept of conversational thread to gain more information and improve the search result in microblog. We also use SVMRank to train a model to determine the rank of relevance of the queries and messages. Through a series of experiments, we proved that our method is easy to implement, and can improve the precision up to 29% in average.