MISCIBILITY AND MORPHOLOGY OF THIN FILMS OF BLENDS OF POLYSTYRENE WITH BROMINATED POLYSTYRENES： EFFECTS OF VARYING THE MOLECULAR WEIGHT, BROMINATION DEGREE AND ANNEALING

Thin films of incompatible polymer blends can form a variety of structures during preparation and subsequent annealing process. For the polymer blend system consisting of polystyrene and poIy(styrene-co-p-bromo-styrene), i.e., PS/PBrxS, its compatibility could be adjusted by varying the degree of bromination and the molecular weight of both components comprised, in this paper, surface chemical compositions of the cast and the annealing films were investigated by X-ray photoelectron spectroscopy (XPS) and contact angle measurement; meanwhile, surface topographical changes are followed by atomic force microscopy (AFM). In addition, substantial attention was paid to the effect of annealing on the morphologic variations induced by phase separation and/or dewetting of the thin film. Moreover, the influences of the molecular weight, Aw, as well as the brominated degree, x%, on the sample surface are explored systematically, and the corresponding observations are explained in virtue of the Flory-Huggins theory, along with the dewetting of the polymer thin film.     

投资者概念关注对股票收益的影响研究—基于百度搜索数据

我国股票市场存在明显的概念炒作现象,扭曲了市场的良性定价机制。本文基于有限注意视角揭示投资者概念关注对股票收益的影响机制,使用百度搜索数据衡量投资者概念关注,以“一带一路”、“5G”和“PM2.5”概念板块的股票为研究样本进行实证研究。结果表明:在控制了Fama-French三因素以及投资者个股关注情况下,投资者概念关注对所属概念板块的股票收益有显著正向影响,但这种影响作用会在随后几期反转为负向影响。投资者个股关注一方面会强化投资者概念关注对股票收益的正向影响,另一方面投资者个股关注越高,个股的流动性就越强,从而会弱化投资者概念关注对股票收益影响作用的反转效应。研究结论丰富了投资者关注的理论研究,为抑制概念炒作提供理论依据和实现途径。     

Learning of shared attention in sociable robotics

Sociable robots are embodied agents that are part of a heterogeneous society of robots and humans. They should be able to recognize human beings and each other, and to engage in social interactions. The use of a robotic architecture may strongly reduce the time and effort required to construct a sociable robot. Such architecture must have structures and mechanisms to allow social interaction, behavior control and learning from environment. Learning processes described on Science of Behavior Analysis may lead to the development of promising methods and structures for constructing robots able to behave socially and learn through interactions from the environment by a process of contingency learning. In this paper, we present a robotic architecture inspired from Behavior Analysis. Methods and structures of the proposed architecture, including a hybrid knowledge representation, are presented and discussed. The architecture has been evaluated in the context of a nontrivial real problem: the learning of the shared attention, employing an interactive robotic head. The learning capabilities of this architecture have been analyzed by observing the robot interacting with the human and the environment. The obtained results show that the robotic architecture is able to produce appropriate behavior and to learn from social interaction.     

PLS-DA Model for the Evaluation of Attention Deficit and Hyperactivity Disorder in Children and Adolescents through Blood Serum FTIR Spectra

Attention deficit and hyperactivity disorder (ADHD) is one of the most common neurodevelopmental disorders of childhood. It affects ~10% of the world’s population of children, and about 30–50% of those diagnosed in childhood continue to show ADHD symptoms later, with 2–5% of adults having the condition. Current diagnosis of ADHD is based on the clinical evaluation of the patient, and on interviews performed by clinicians with parents and teachers of the children, which, together with the fact that it shares common symptoms and frequent comorbidities with other neurodevelopmental disorders, makes the accurate and timely diagnosis of the disorder a difficult task. Despite the large effort to identify reliable biomarkers that can be used in a clinical environment to support clinical diagnosis, this goal has never been achieved hitherto. In the present study, infrared spectroscopy was used together with multivariate statistical methods (hierarchical clustering and partial least-squares discriminant analysis) to develop a model based on the spectra of blood serum samples that is able to distinguish ADHD patients from healthy individuals. The developed model used an approach where the whole infrared spectrum (in the 3700–900 cm⁻¹ range) was taken as a holistic imprint of the biochemical blood serum environment (spectroscopic biomarker), overcoming the need for the search of any particular chemical substance associated with the disorder (molecular biomarker). The developed model is based on a sensitive and reliable technique, which is cheap and fast, thus appearing promising to use as a complementary diagnostic tool in the clinical environment.     

基于神经网络的股票预测模型

针对股票市场的特征提取困难、预测精度较低等问题,本文基于深度学习算法,构建了一系列用于股票市场预测的神经网络模型,包括基于多层感知机(MLP)、卷积神经网络(CNN)、递归神经网络(RNN)、长短期记忆网络(LSTM)和门控神经单元(GRU)的模型。 针对RNN、LSTM和GRU无法充分利用所参考的时间维度的信息,引入注意力机制(Attention Mechanism) 给各时间维度的信息赋予不同权重,区分不同信息对预测的重要程度,从而提升递归网络模型的性能。上述模型均基于股票数据进行了优化,基于上证指数对各类模型进行了充分的对比实验,探索了模型中重要变量对性能的影响,旨在为基于神经网络的股票预测模型给出具体的优化方向。     

Part-Aware Mask-Guided Attention for Thorax Disease Classification

Thorax disease classification is a challenging task due to complex pathologies and subtle texture changes, etc. It has been extensively studied for years largely because of its wide application in computer-aided diagnosis. Most existing methods directly learn global feature representations from whole Chest X-ray (CXR) images, without considering in depth the richer visual cues lying around informative local regions. Thus, these methods often produce sub-optimal thorax disease classification performance because they ignore the very informative pathological changes around organs. In this paper, we propose a novel Part-Aware Mask-Guided Attention Network (PMGAN) that learns complementary global and local feature representations from all-organ region and multiple single-organ regions simultaneously for thorax disease classification. Specifically, multiple innovative soft attention modules are designed to progressively guide feature learning toward the global informative regions of whole CXR image. A mask-guided attention module is designed to further search for informative regions and visual cues within the all-organ or single-organ images, where attention is elegantly regularized by automatically generated organ masks and without introducing computation during the inference stage. In addition, a multi-task learning strategy is designed, which effectively maximizes the learning of complementary local and global representations. The proposed PMGAN has been evaluated on the ChestX-ray14 dataset and the experimental results demonstrate its superior thorax disease classification performance against the state-of-the-art methods.     

媒体信息量异常或有害股票收益

本文借助一个独特的数据样本,运用媒体对股票的剩余关注度模型,实证研究异常媒体信息量与股票收益之间的关系,以期为投资者进行投资决策提供一定的参考和指导。研究发现:异常媒体信息量越大,该股票在下一个月的平均收益率越低,存在媒体效应;由此所构造的零投资组合经CAPM模型、FF三因素模型和Car-hart四因素模型调整后,均能获取显著的超额收益,结果具有稳健性。此外,实证结果还表明媒体效应所带来的超额收益源于媒体信息量异常大的股票组合的显著低收益,本文认为,这种不对称现象产生的原因可能更多的是由投资者情绪导致的股票价格对媒体报道的过度反应,并进而导致较低的期望收益。     

基于遗传注意力机制的DLSTM电力系统混沌预测

提出一种基于遗传算法优化注意力机制的深度长短期记忆网络(DLSTM)方法,用于电力系统的混沌预测。通过传递共享参数,将遗传算法优化的注意力机制加入DLSTM模型中,可以挖掘时间序列中潜在特征,同时避免陷入局部优化。该方法是一种受进化计算方法启发的寻优方法,可以很好地学习注意力层中的参数。电力系统混沌预测实验表明所提模型比其他参考模型具有更高的预测精度和长期预测能力。     

基于注意力和角度间隔损失的高光谱目标跟踪北大核心CSCD

随着计算机技术的发展,基于深度学习的目标跟踪方法已成为计算机视觉领域中重要的研究方向;但跟踪环境的复杂多变使得跟踪算法在背景干扰、颜色相近等问题上仍面临巨大挑战。相比于传统彩色图像,高光谱图像包含丰富的辐射、空间和光谱信息,能够有效提升目标跟踪的准确率。提出了将注意力机制(attention mechanism)和加性角度间隔损失(additive angular margin loss, AAML)相结合的方法来进行针对高光谱图像的目标跟踪。通过融合多域神经网络对不同波段组合进行特征提取,同时设计了融合的注意力机制模型,使得来自不同波段组合之间的相似特征进行整合和强化,在目标背景颜色相近的情况下,网络会更多地注意目标物体,使得跟踪结果更为准确。在此基础上为了使目标和背景的区分更具有判别性,网络使用加性角度间隔损失作为损失函数,在训练过程中可以有效减小同类样本的类内距离,增大正负类样本的类间距离,从而提高网络的准确性和稳定性。实验结果表明,本文方法可使两种跟踪精度评价指标精确率和成功率分别提升1.3%和0.3%,相较于其他方法更具优势。     

fMRI assessment of thalamocortical connectivity during attentional performance

Functional magnetic resonance imaging (fMRI) studies have shown dysfunction in key areas associated with the thalamocortical circuit in patients with schizophrenia. This study examined the functional connectivity involving the frontal-thalamic circuitry during a spatial focusing-of-attention task in 18 unmedicated patients with schizophrenia and 38 healthy controls. Functional connectivity was analyzed by assigning seed regions (in the thalamic nuclei (mediodorsal nucleus (MDN), pulvinar, anterior nucleus (AN)), the dorsolateral prefrontal cortex (Brodmann areas 9 and 46), and the caudate), and correlating their respective activity with that in the non-seed regions voxel-wise. Functional connectivity analysis demonstrated that functional connectivity was significantly impaired in patients, e.g., between the right pulvinar and regions such as the prefrontal and temporal cortices and the cerebellum. On the other hand, enhanced functional connectivity was found in patients e.g., between the AN and regions such as the prefrontal and temporal cortices. In addition, the patients had significantly lower task performance and less (but non-significant) brain activation than those of controls. These results revealed disturbed functional integration in schizophrenia, and suggested that the functional connectivity abnormalities in the thalamocortical circuitry, especially the frontal-thalamic circuitry, may underlie the attention deficits in schizophrenia patients. Further, this study suggested that functional connectivity analysis might be more sensitive than brain activation analysis in detecting the functional abnormalities in schizophrenia.