基于双向循环神经网络的汉语语音识别*

当前基于深度神经网络模型中,虽然其隐含层可设置多层,对复杂问题适应能力强,但每层之间的节点连接是相互独立的,这种结构特性导致了在语音序列中无法利用上下文相关信息来提高识别效果,而传统的循环神经网络虽然做出了改进,但是只能对上文信息进行利用。针对以上问题,该文采用可以同时利用语音序列中上下文相关信息的双向循环神经网络模型与深度神经网络模型相结合,并应用于语音识别。构建具有5层隐含层的模型,其中第3层为双向循环神经网络结构,其他层采用深度神经网络结构。实验结果表明:加入了双向循环神经网络结构的模型与其他模型相比,较好地提高了识别正确率;噪声对双向循环神经网络汉语识别有重要影响,尤其是训练集和测试集附加噪声类型不同时,单一的含噪声语音的训练模型无法适应不同噪声类型的语音识别;调整神经网络模型中隐含层神经元数量后,识别正确率并不是一直随着隐含层中神经元数量的增加而增加,神经元数量数目增加到一定程度后正确率出现了降低的趋势。     

高中物理“教-学-评一体化”课堂教学探索——以“示波器的原理——带电粒子在电场中的偏转”为例

"教-学-评一体化"的课堂是指围绕教学目标,教师的教学、学生的学习以及教师对学生的评价组成一个有机的、整体的有效课堂.教学"有效"的唯一证据在于目标的达成,在于学生学习结果的质量,在于何以证明学生学会了什么.因此,教学中要关注对学生的评价.本文以"示波器的原理——带电粒子在电场中的偏转"为例,论述在"教-学-评一体化"的课堂中如何用评价促进学生思维发展.     

A survey on HHL algorithm: From theory to application in quantum machine learning

The Harrow-Hassidim-Lloyd (HHL) algorithm is a method to solve the quantum linear system of equations that may be found at the core of various scientific applications and quantum machine learning models including the linear regression, support vector machines and recommender systems etc. After reviewing the necessary background on elementary quantum algorithms, we provide detailed account of how HHL is exploited in different quantum machine learning (QML) models, and how it provides the desired quantum speedup in all these models. At the end, we briefly discuss some of the remaining challenges ahead for HHL-based QML models and related methods.     

相位可控的核四极矩共振激励脉冲发生器设计

在核四极矩共振（NQR）领域,射频激励脉冲信号的优劣对NQR响应信号有重要影响.针对常规方法中射频激励脉冲参数不可控的问题,本文基于32位闪存微型控制器STM32和直接数字频率合成（DDS）芯片AD9910设计了一种相位可控激励脉冲发生器.采用STM32控制AD9910产生波形参数（脉冲宽度、脉冲间隔、脉冲个数和共振频率等）可控的射频激励脉冲,利用LabVIEW软件平台设计脉冲参数设置界面,并建立计算机与微控制器通信,实现波形参数的精确优化控制.实验结果表明,该方法实现了相位可控的NQR激励脉冲序列,可为后续NQR信号检测提供有效激励源.     

Adaptive Neuro-Fuzzy Inference System and a Multilayer Perceptron Model Trained with Grey Wolf Optimizer for Predicting Solar Diffuse Fraction

The accurate prediction of the solar diffuse fraction (DF), sometimes called the diffuse ratio, is an important topic for solar energy research. In the present study, the current state of Diffuse irradiance research is discussed and then three robust, machine learning (ML) models are examined using a large dataset (almost eight years) of hourly readings from Almeria, Spain. The ML models used herein, are a hybrid adaptive network-based fuzzy inference system (ANFIS), a single multi-layer perceptron (MLP) and a hybrid multi-layer perceptron grey wolf optimizer (MLP-GWO). These models were evaluated for their predictive precision, using various solar and DF irradiance data, from Spain. The results were then evaluated using frequently used evaluation criteria, the mean absolute error (MAE), mean error (ME) and the root mean square error (RMSE). The results showed that the MLP-GWO model, followed by the ANFIS model, provided a higher performance in both the training and the testing procedures.     

Semiotic Aggregation in Deep Learning

Convolutional neural networks utilize a hierarchy of neural network layers. The statistical aspects of information concentration in successive layers can bring an insight into the feature abstraction process. We analyze the saliency maps of these layers from the perspective of semiotics, also known as the study of signs and sign-using behavior. In computational semiotics, this aggregation operation (known as superization) is accompanied by a decrease of spatial entropy: signs are aggregated into supersign. Using spatial entropy, we compute the information content of the saliency maps and study the superization processes which take place between successive layers of the network. In our experiments, we visualize the superization process and show how the obtained knowledge can be used to explain the neural decision model. In addition, we attempt to optimize the architecture of the neural model employing a semiotic greedy technique. To the extent of our knowledge, this is the first application of computational semiotics in the analysis and interpretation of deep neural networks.     

Change in motor cortex activation for muscle release by motor learning

For central nervous system disorders'' rehabilitation, it is important to accurately understand motor control and implement an appropriate motor learning process to induce neuroplastic changes. The neurophysiological studies have revealed that neural control mechanisms are crucial during both the onset of muscular activities and muscle release after contraction. When performing various movements during daily activities, muscle relaxation control enables precise force output and timing control. Moreover, surround inhibition is a functional mechanism in the motor system. Surround inhibition of the motor system may be involved in the selective execution of desired movements. This review demonstrates cortical excitability resulting from motor learning, movement control mechanisms including muscle relaxation and the suppression of nontarget muscle groups, and the voluntary drive''s importance that is required for movement.     

利用DSP底层结构提高MPEG-4编码的实时性

由于利用TMS320C6000系列的DSP底层结构可以提高图像压缩的实时性,本文在TMS320C6416平台上针对DSP底层结构对MPEG-4视频编码进行了优化,包括对8个运算单元能并行执行的特性进行汇编。针对DSP芯片两级高速缓存的工作特点,对编码器中代码和数据所占用的内存空间进行优化。利用EDMA级联特性在内存开辟双缓冲区,同时完成视频数据编码和传输的工作,从而有效地改善编码效率。实验结果表明：该编码器可以对512 pixel×512 pixel大小的灰度图像,以30 frame/s帧频进行压缩,实现了对视频图像的实时编码。     

Digital pulse shape discrimination methods for n-γ separation in an EJ-301 liquid scintillation detector

A digital pulse shape discrimination system based on a programmable module NI-5772 has been established and tested with an EJ-301 liquid scintillation detector. The module was operated by running programs developed in Lab VIEW, with a sampling frequency up to 1.6 GS/s. Standard gamma sources22 Na,137Cs and60 Co were used to calibrate the EJ-301 liquid scintillation detector, and the gamma response function was obtained. Digital algorithms for the charge comparison method and zero-crossing method have been developed. The experimental results show that both digital signal processing(DSP) algorithms can discriminate neutrons from γ-rays. Moreover,the zero-crossing method shows better n-γ discrimination at 80 ke Vee and lower, whereas the charge comparison method gives better results at higher thresholds. In addition, the figure-of-merit(FOM) for detectors of two different dimensions were extracted at 9 energy thresholds, and it was found that the smaller detector presented better n-γseparation for fission neutrons.     

大力开展我国电磁兼容技术的科研工作及若干建议

：本文介绍了国际上关于环境电磁学及电磁兼容技术的一些科研进展情况,描述了电磁兼容的动态,最后并提出了若干建议.