Aspect-based sentiment classification is a popular task aimed at identifying the corresponding emotion of a specific aspect. One sentence may contain various sentiments for different aspects. Many sophisticated methods such as attention mechanism and Convolutional Neural Networks (CNN) have been widely employed for handling this challenge. Recently, semantic dependency tree implemented by Graph Convolutional Networks (GCN) is introduced to describe the inner connection between aspects and the associated emotion words. But the improvement is limited due to the noise and instability of dependency trees. To this end, we propose a dependency graph enhanced dual-transformer network (named DGEDT) by jointly considering the flat representations learnt from Transformer and graphbased representations learnt from the corresponding dependency graph in an iterative interaction manner. Specifically, a dualtransformer structure is devised in DGEDT to support mutual reinforcement between the flat representation learning and graph-based representation learning. The idea is to allow the dependency graph to guide the representation learning of the transformer encoder and vice versa. The results on five datasets demonstrate that the proposed DGEDT outperforms all state-of-the-art alternatives with a large margin.

No abstract

Pre-harvest sprouting (PHS) is mainly caused by the breaking of seed dormancy in high rainfall regions, which leads to huge economic losses in wheat. In this study, we evaluated 717 Chinese wheat landraces for PHS resistance and carried out genome-wide association studies (GWAS) using to 9,740 DArT-seq and 178,803 SNP markers. Landraces were grown across six environments in China and germination testing of harvest-ripe grain was used to calculate the germination rate (GR) for each accession at each site. GR was highly correlated across all environments. A large number of landraces (194) displayed high levels of PHS resistance (i.e., mean GR < 0.20), which included nine white-grained accessions. Overall, white-grained accessions displayed a significantly higher mean GR (42.7–79.6%) compared to red-grained accessions (19.1–56.0%) across the six environments. Landraces from mesic growing zones in southern China showed higher levels of PHS resistance than those sourced from xeric areas in northern and north-western China. Three main quantitative trait loci (QTL) were detected by GWAS: one on 5D that appeared to be novel and two co-located with the grain color transcription factor Tamyb10 on 3A and 3D. An additional 32 grain color related QTL (GCR-QTL) were detected when the set of red-grained landraces were analyzed separately. GCR-QTL occurred at high frequencies in the red-grained accessions and a strong correlation was observed between the number of GCR-QTL and GR (R2 = 0.62). These additional factors could be critical for maintaining high levels of PHS resistance and represent targets for introgression into white-grained wheat cultivars. Further, investigation of the origin of haplotypes associated with the three main QTL revealed that favorable haplotypes for PHS resistance were more common in accessions from higher rainfall zones in China. Thus, a combination of natural and artificial selection likely resulted in landraces incorporating PHS resistance in China.

Pulmonary nodule detection, false positive reduction and segmentation represent three of the most common tasks in the computer aided analysis of chest CT images. Methods have been proposed for each task with deep learning based methods heavily favored recently. However training deep learning models to solve each task separately may be sub-optimal -resource intensive and without the benefit of feature sharing. Here, we propose a new end-to-end 3D deep convolutional neural net (DCNN), called NoduleNet, to solve nodule detection, false positive reduction and nodule segmentation jointly in a multi-task fashion. To avoid friction between different tasks and encourage feature diversification, we incorporate two major design tricks: 1) decoupled feature maps for nodule detection and false positive reduction, and 2) a segmentation refinement subnet for increasing the precision of nodule segmentation. Extensive experiments on the large-scale LIDC dataset demonstrate that the multi-task training is highly beneficial, improving the nodule detection accuracy by 10.27%, compared to the baseline model trained to only solve the nodule detection task. We also carry out systematic ablation studies to highlight contributions from each of the added components. Code is available at https://github.com/uci-cbcl/NoduleNet.

Early detection of pulmonary nodules in computed tomography (CT) images is essential for successful outcomes among lung cancer patients. Much attention has been given to deep convolutional neural network (DCNN)-based approaches to this task, but models have relied at least partly on 2D or 2.5D components for inherently 3D data. In this paper, we introduce a novel DCNN approach, consisting of two stages, that is fully three-dimensional end-to-end and utilizes the state-of-the-art in object detection. First, nodule candidates are identified with a U-Net-inspired 3D Faster R-CNN trained using online hard negative mining. Second, false positive reduction is performed by 3D DCNN classifiers trained on difficult examples produced during candidate screening. Finally, we introduce a method to ensemble models from both stages via consensus to give the final predictions. By using this framework, we ranked first of 2887 teams in Season One of Alibaba's 2017 TianChi AI Competition for Healthcare.

No abstract

Simple sequence repeats (SSR) – also known as microsatellites – have been used extensively in genetic analysis, fine mapping, quantitative trait locus (QTL) mapping, as well as marker-assisted selection (MAS) breeding and other techniques. Despite a plethora of studies reporting that perfect SSRs with stable motifs and flanking sequences are more efficient for genetic research, the lack of a high throughput technology for SSR genotyping has limited their use as genetic targets in many crops. In this study, we developed a technology called Target SSR-seq that combined the multiplexed amplification of perfect SSRs with high throughput sequencing. This method can genotype plenty of SSR loci in hundreds of samples with highly accurate results, due to the substantial coverage afforded by high throughput sequencing. We also detected 844 perfect SSRs based on 182 resequencing datasets in cucumber, of which 91 SSRs were selected for Target SSR-seq. Finally, 122 SSRs, including 31 SSRs for varieties identification, were used to genotype 382 key cucumber varieties readily available in Chinese markets using our Target SSR-seq method. Libraries of PCR products were constructed and then sequenced on the Illumina HiSeq X Ten platform. Bioinformatics analysis revealed that 111 filtered SSRs were accurately genotyped with an average coverage of 1289× at an extremely low cost; furthermore, 398 alleles were observed in 382 cucumber cultivars. Genetic analysis identified four populations: northern China type, southern China type, European type, and Xishuangbanna type. Moreover, we acquired a set of 16 core SSRs for the identification of 382 cucumber varieties, of which 42 were isolated as backbone cucumber varieties. This study demonstrated that Target SSR-seq is a novel and efficient method for genetic research.

In the context of neuro-pathological disorders, neuroimaging has been widely accepted as a clinical tool for diagnosing patients with Alzheimer’s disease (AD) and mild cognitive impairment (MCI). The advanced deep learning method, a novel brain imaging technique, was applied in this study to evaluate its contribution to improving the diagnostic accuracy of AD. Three-dimensional convolutional neural networks (3D-CNNs) were applied with magnetic resonance imaging (MRI) to execute binary and ternary disease classification models. The dataset from the Alzheimer’s disease neuroimaging initiative (ADNI) was used to compare the deep learning performances across 3D-CNN, 3D-CNN-support vector machine (SVM) and two-dimensional (2D)-CNN models. The outcomes of accuracy with ternary classification for 2D-CNN, 3D-CNN and 3D-CNN-SVM were [Formula: see text]%, [Formula: see text]% and [Formula: see text]% respectively. The 3D-CNN-SVM yielded a ternary classification accuracy of 93.71%, 96.82% and 96.73% for NC, MCI and AD diagnoses, respectively. Furthermore, 3D-CNN-SVM showed the best performance for binary classification. Our study indicated that ‘NC versus MCI’ showed accuracy, sensitivity and specificity of 98.90%, 98.90% and 98.80%; ‘NC versus AD’ showed accuracy, sensitivity and specificity of 99.10%, 99.80% and 98.40%; and ‘MCI versus AD’ showed accuracy, sensitivity and specificity of 89.40%, 86.70% and 84.00%, respectively. This study clearly demonstrates that 3D-CNN-SVM yields better performance with MRI compared to currently utilized deep learning methods. In addition, 3D-CNN-SVM proved to be efficient without having to manually perform any prior feature extraction and is totally independent of the variability of imaging protocols and scanners. This suggests that it can potentially be exploited by untrained operators and extended to virtual patient imaging data. Furthermore, owing to the safety, noninvasiveness and nonirradiative properties of the MRI modality, 3D-CNN-SMV may serve as an effective screening option for AD in the general population. This study holds value in distinguishing AD and MCI subjects from normal controls and to improve value-based care of patients in clinical practice.