基于BP神经网络的增益量化中的阶距自适应

本文分析了G.728语音编码算法增益固定量化的缺陷,设计了精确表示增益的自适应量化方案。采用BP神经网络对增益的量化步长进行非线性自适应量化,并提出一种新的用于训练BP神经网络的数据的获取方法。     

大学生学习焦虑心理研究

通过分析大学生学习焦虑心理的成因以及学习焦虑控制问题。论述了大学生学习焦虑心理与学习成绩的关系。     

Using genetically-defined rodent strains for the identification of hippocampal traits relevant for two-way avoidance behavior: a non-invasive approach

Summary Genetically-defined rodent strains permit the identification of hippocampal traits which are of functional relevance for the performance of two-way avoidance behavior. This is exemplified here by analyzing the relationship between infrapyramidal mossy fibers (a tiny projection terminating upon the basal dendrites of hippocampal pyramidal neurons) and two-way avoidance learning in about 800 animals. The necessary steps include 1) identification of structural traits sensitive to selective breeding for extremes in two-way avoidance, 2) testing the robustness of the associations found by studying individual and genetical correlations between hippocampal traits and behavior, 3) establishing causal relationships by Mendelian crossing of strains with extreme structural traits and studying the behavioral consequences of such structural randomization, 4) confirming causal relationships by manipulating the structural variable in inbred (isogenic) strains, thereby eliminating the possibility of genetic linkage, and 5) ruling out the possibility of spurious associations by studying the correlations between the hippocampal trait and other behaviors known to depend on hippocampal functioning.In comparison with the classical lesion approach for identifying relationships between brain and behavior, the present procedure appears to be superior in two aspects: it is non-invasive, and it focuses automatically on those brain traits which are used by natural selection to shape behaviorally-defined animal populations, i.e., it reveals the natural regulators of behavior.     

Learning attention-based strategies to cooperate for multi-agent path finding

Multi-agent path finding (MAPF) is a challenging multi-agent systems problem where all agents are required to effectively reach their goals concurrently with not colliding with each other and avoiding obstacles. In MAPF, it is a challenge to effectively express the observation of agents, utilize historical information, and effectively communicate with neighbor agents. To tackle these issues, in this work, we proposed a well-designed model that utilizes the local states of nearby agents and outputs an optimal action for each agent to execute. We build the local observation encoder by using residual attention CNN to extract local observations and use the Transformer architecture to build an interaction layer to combine local observations of agents. With the purpose of overcoming the deficiency of success rate, we also designed a new evaluation index, namely extra time rate (ETR). The experimental results show that our model is superior to most previous models in terms of success rate and ETR. In addition, we also completed the ablation study on the model, and the effectiveness of each component of the model was proved.     

基于强化学习的电磁悬浮型磁浮列车悬浮控制

为了保证磁浮列车的安全、可靠运行,研究了悬浮系统在参数摄动条件下的悬浮控制问题。首先,对电磁悬浮（EMS）型磁浮列车的基本悬浮单元建模,给出了电流控制模型;然后,建立了悬浮系统的强化学习环境以及软演员-评论家（SAC）智能体,并设计了加速训练的奖励函数与“吸死”处理方案;最后,提出了基于强化学习的悬浮控制方法。与传统比例-积分-微分（PID）控制方法的对比结果表明,本方法具有更快的动态响应,在损失50%线圈匝数或磁极面积变化时具有更好的跟踪精度。     

多尺度自适应Gamma矫正的低照图像增强

为了在提高低照图像的亮度和对比度的同时,保持图像的色彩自然度,提出了多尺度自适应Gamma矫正的低照图像增强方法.鉴于HSI颜色空间的明度分量Ⅰ决定着图像的明暗程度,以及明度分量的多尺度特性,首先,将图像转换到HSI颜色空间,对明度分量Ⅰ进行多尺度Retinex分解;然后,对分解得到的多尺度光照图像分别进行自适应的Gamma矫正,其中Gamma指数自适应于暗区像素的占比,鲁棒地改善光照图像的光照分布,经Retinex反变换得到增强的多尺度明度分量;最后,将增强的多尺度明度分量的线性融合作为增强的明度分量Ⅰ’,与色相分量H和饱和度分量S重组,得到最后的增强图像.实验结果表明,相对于部分最新提出的现有方法,新方法的PSNR和SSIM值分别高出1.19 dB和1.8%以上,具有更好的增强效果和鲁棒性,增强图像的视觉效果更适宜,对比度更高.     

基于深度学习的红外与可见光图像融合综述

图像融合技术是指从不同的源图像中提取并融合互补的信息,生成一幅信息量更丰富、对后续高级视觉任务提供足够支持的图像.红外与可见光图像融合（Infrared and Visible Image Fusion,IVIF）是图像融合领域的一个重要分支.近年来,深度学习技术在视觉计算领域表现出了良好的性能,尤其是基于自编码器、卷积神经网络、生成对抗网络等几种基于深度学习的IVIF技术得到了蓬勃发展.为此,对基于深度学习的IVIF算法的方法、数据集和评估指标等进行了总结和阐述;通过大量的实验,进行定性和定量的结果分析,对比了各类基于深度学习IVIF算法的性能;最后,讨论了该领域未来发展的一些前景和研究方向.     

联合实体边界检测的命名实体识别方法

针对传统命名实体识别方法无法有效利用实体边界信息的问题,提出一种联合实体边界检测的命名实体识别方法,即将实体边界检测作为辅助任务,增强模型对实体边界的判断能力,进而提升模型对实体的识别效果。首先,利用Bert预训练语言模型对原始文本进行特征嵌入获取词向量,并引入自注意力机制增强词对上下文信息的利用;其次,在命名实体识别任务的基础上,添加实体边界检测辅助任务,增强模型对实体边界的识别能力;再次,对比联合实体边界检测的命名实体识别方法与基线方法的有效性,并对测试结果进行消融实验;最后,进行样例分析,分析损失权重β对实体边界检测的影响。实验结果表明,在英文社交媒体数据集Twitter-2015上,联合实体边界检测的命名实体识别方法相较于基线模型取得了更高的精准率、召回率和F1值,其中F1值达到了73.57%;并且,边界检测辅助任务提升了基线方法的检测效果。所提方法能有效利用实体边界信息,从而获得更好的实体识别效果,促进了人机交互系统的发展,对自然语言处理下游任务有重要意义。     

数据驱动下交通异常行为的双层识别模型

为实现对车辆异常行为的准确识别,提高车辆行驶的安全性,提出了一种基于支持向量机(support vector machine, SVM)、监督学习与长短期记忆(long short term memory, LSTM)深度学习的交通异常驾驶行为双层识别模型。首先,对车辆轨迹数据筛除和滤波,构建异常行为数据集;其次,从异常行为轨迹特征中提取出特定异常行为的特征标签,并人为标定在训练集中;再次,构建SVM模型对训练集进行粗识别,基于SVM的二分法原理,从测试集中筛选出异常行为;最后,通过LSTM时间序列模型构建具体种类的异常行为模型,并通过深度学习的方法,从异常行为数据中细分为蛇形驾驶、急速变向、侧滑、大半径转弯、快速U型转弯、急刹车等具体的异常驾驶行为。本次实验选用下一代仿真(next generation simulation, NGSIM)数据中US-101高速公路和peachtree城市道路的数据集的轨迹数据验证SVM和LSTM双层识别模型的性能,包括均方根误差、识别准确率等。结果表明,构建的双层识别模型在第一层有98%的识别准确率,第二层有超过80%的识别准确率,可以较为准确地识...     

Hypothalamic histamine modulates adaptive behavior of rats at high environmental temperature

Summary Histamine content in the rat hypothalamus was lower at 4°C and higher at 31°C compared to that at 21°C. Pretreatment with -fluoromethylhistidine, a suicide inhibitor of histidine decarboxylase, attenuated both the increased level of hypothalamic histamine and rat adaptive behavior at 31°C. Increase of histamine content in the hypothalamus appears to be an important factor contributing to rat adaptive behavior to high environmental temperature.