An unsupervised approach to generating generic summaries of documents

We present an optimization-based unsupervised approach to automatic document summarization. In the proposed approach, text summarization is modeled as a Boolean programming problem. This model generally attempts to optimize three properties, namely, (1) relevance: summary should contain informative textual units that are relevant to the user; (2) redundancy: summaries should not contain multiple textual units that convey the same information; and (3) length: summary is bounded in length. The approach proposed in this paper is applicable to both tasks: single- and multi-document summarization. In both tasks, documents are split into sentences in preprocessing. We select some salient sentences from document(s) to generate a summary. Finally, the summary is generated by threading all the selected sentences in the order that they appear in the original document(s). We implemented our model on multi-document summarization task. When comparing our methods to several existing summarization methods on an open DUC2005 and DUC2007 data sets, we found that our method improves the summarization results significantly. This is because, first, when extracting summary sentences, this method not only focuses on the relevance scores of sentences to the whole sentence collection, but also the topic representative of sentences. Second, when generating a summary, this method also deals with the problem of repetition of information. The methods were evaluated using ROUGE-1, ROUGE-2 and ROUGE-SU4 metrics. In this paper, we also demonstrate that the summarization result depends on the similarity measure. Results of the experiment showed that combination of symmetric and asymmetric similarity measures yields better result than their use separately.     

When danger lurks in the background: Attentional capture by animal fear-relevant distractors is specific and selectively enhanced by animal fear.

Across 2 experiments, a new experimental procedure was used to investigate attentional capture by animal fear-relevant stimuli. In Experiment 1 (N = 34), unselected participants were slower to detect a neutral target animal in the presence of a spider than a cockroach distractor and in the presence of a snake than a large lizard distractor. This result confirms that phylogenetically fear-relevant animals capture attention specifically and to a larger extent than do non-fear-relevant animals. In Experiment 2 (N = 86), detection of a neutral target animal was slowed more in the presence of a feared fear-relevant distractor (e.g., a snake for snake-fearful participants) than in presence of a not-feared fear-relevant distractor (e.g., a spider for snake-fearful participants). These results indicate preferential attentional capture that is specific to phylogenetically fear-relevant stimuli and is selectively enhanced in individuals who fear these animals. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

折半查找算法在电子会议系统调试中的应用

关联性故障代表单元的查找定位是电子会议系统调试的关键环节。为了提高查找定位效率,提出了一种基于折半查找的快速定位算法。通过对电子会议系统中干线上的代表单元可靠度的分析,把代表单元的故障情况转化成随着代表单元数量动态变化的量化数据,构成一组有序的可靠度数据,将工程调试问题转化成数据分析问题,引入折半查找算法,设计高效准确定位故障代表单元的算法,并进行了算法效率分析。该方法操作简单,无需借助特殊工具即可应用于复杂环境下电子会议系统的现场调试。     

基于LASVM-NC和TF.RF的文本分类方法

非凸在线支持向量机（LASVM-NC）具有抗噪能力强和训练速度快的优点,而词频相关频率积（tf.rf）则是一种自适应能力很强、分类性能非常好的文本特征。通过把非凸在线支持向量机和词频相关频率积相结合,提出了一种新的文本分类方法,即LASVM-NC＋tf.rf。实验结果表明,这种方法在LASVM-NC与多种其他特征的结合中性能是最好的,且与SVM＋tf.rf相比,不仅所产生的分类器具有泛化能力更强、模型表达更稀疏的优点,而且在处理含噪声的数据时具有更好的鲁棒性,在处理大规模数据时具有快得多的训练速度。     

基于相关向量机的MEMS陀螺仪随机漂移补偿

提出了一种基于相关向量机的MEMS陀螺仪随机漂移预测方法。针对MEMS陀螺仪随机漂移误差的非线性、不确定性等特点,建立了相关向量机预测模型,并采用EM算法获得模型的参数。针对随机漂移的混沌特性,利用相空间重构技术,将重构后的漂移序列作为输入变量进行模型的训练和预测。训练和测试结果均表明,该方法具有很好的预测效果,优于常用的时间序列分析法和支持向量机法。利用预测结果对随机漂移进行补偿,有效地提高了陀螺的使用精度。     

基于RNN-RBM语言模型的语音识别研究

近年来深度学习兴起,其在语言模型领域有着不错的成效,如受限玻尔兹曼机(restricted Boltzmann machine,RBM)语言模型等.不同于N-gram语言模型,这些根植于神经网络的语言模型可以将词序列映射到连续空间来评估下一词出现的概率,以解决数据稀疏的问题.此外,也有学者使用递归神经网络来建构语言模型,期望由递归的方式充分利用所有上文信息来预测下一词,进而有效处理长距离语言约束.根据递归受限玻尔兹曼机神经网络(recurrent neural network-restricted Boltzmann machine,RNN-RBM)的基础来捕捉长距离信息;另外,也探讨了根据语言中语句的特性来动态地调整语言模型.实验结果显示,使用RNN-RBM语言模型对于大词汇连续语音识别的效能有相当程度的提升.     

基于M-Chord的分布式视觉监控交互检索算法

为对城域视觉监控系统产生的海量多媒体数据进行高效索引和检索,提出基于M-Chord的分布式交互检索和负载均衡算法.对象以颜色柱状图、Gabor纹理描述和轨迹系数空间特征组成的特征向量描述,由M-Chord算法进行分布式索引分发和检索.提出基于线性和非线性支持向量机的分布式相关反馈算法,将支持向量机模型建立的度量空间与M-Chord索引所处的度量空间结合,利用分片中心点减少了对节点和对象的访问.实验表明提出的算法可以准确反映用户的查询语义,仅需要检索分布式监控网络中的少部分节点即可达到与顺序扫描接近的查准率,同时较好实现了分布式监控网络中的负载均衡.     

自适应多核组合相关向量机预测方法及其在机械设备剩余寿命预测中的应用

针对支持向量机(Support vector machine,SVM)的惩罚系数难以确定、核函数必须满足Mercer定理等问题,相关向量机(Relevance vector machine,RVM)应运而生以解决上述问题,并在趋势预测等领域得到一定的应用。核函数是决定RVM预测精度的关键因素之一,目前的研究通常是人为选择单一核函数,因此增加了对参数的依赖性并降低了RVM预测的鲁棒性。为了解决以上问题,提出一种新的自适应多核组合RVM预测方法。该方法首先选择多个核函数,利用粒子滤波产生核函数权重,建立多核组合RVM集,然后经过不断地迭代预测、权值更新和重采样,自适应获取最优多核组合RVM,从而自适应融合多个核函数的特性,克服基于单一核函数RVM的局限,提高预测精度和鲁棒性。利用仿真对提出方法进行了验证,并将其应用于机械设备的剩余寿命预测,取得了比基于单一核函数RVM更好的预测效果。     

基于多尺度相关向量机的城市日用水量预测

为提高日用水量预测精度,提出一种基于多尺度相关向量机的预测模型。通过静态小波分解将用水量非平稳时间序列分解为不同尺度的平稳时间序列,然后在分解后的各子序列分别建立相关向量机回归模型进行预测,最后通过小波逆变换将各子序列预测结果整合得出原始用水量时间序列的预测值。在实例分析中分别利用多尺度关联向量机模型和单尺度相关向量机预测模型对实际用水量进行预测分析。结果表明,前者具有更高的预测精度,可应用于城市日用水量的预测。     

基于相关向量机的高光谱图像解混方法

针对传统的高光谱数据解混方法中存在的解混精度不高、丰度图模糊的缺陷,提出一种基于相关向量机的高光谱图像解混方法(unmixing algorithm based on relevance vector machine,UARVM)。其核心思想是采用改进的一对余型的相关向量机将多分类问题转化为多个二分类的问题,且求取出每个样本所对应的归属类别的概率值,即丰度值来完成图像的解混。理论研究和仿真结果表明:相对于传统解混方法,UARVM解混精度高,丰度分布图效果好。