机器人听觉声源定位研究综述

声源定位技术定位出外界声源相对于机器人的方向和位置,机器人听觉声源定位系统可以极大地提高机器人与外界交互的能力.总结和分析面向机器人听觉的声源定位技术对智能机器人技术的发展有着重要的意义.首先总结了面向机器人听觉的声源定位系统的特点,综述了机器人听觉声源定位的关键技术,包括到达时间差、可控波束形成、高分辨率谱估计、双耳听觉、主动听觉和视听融合技术.其次对麦克风阵列模型进行了分类,比较了基于三维麦克风阵列、二维麦克风阵列和双耳的7个典型系统的性能.最后总结了机器人听觉声源定位系统的应用,并分析了存在的问题和未来的发展趋势.     

A new algorithm for fast mining frequent itemsets using N-lists

Mining frequent itemsets has emerged as a fundamental problem in data mining and plays an essential role in many important data mining tasks.In this paper,we propose a novel vertical data representation called N-list,which originates from an FP-tree-like coding prefix tree called PPC-tree that stores crucial information about frequent itemsets.Based on the N-list data structure,we develop an efficient mining algorithm,PrePost,for mining all frequent itemsets.Efficiency of PrePost is achieved by the following three reasons.First,N-list is compact since transactions with common prefixes share the same nodes of the PPC-tree.Second,the counting of itemsets’ supports is transformed into the intersection of N-lists and the complexity of intersecting two N-lists can be reduced to O(m + n) by an efficient strategy,where m and n are the cardinalities of the two N-lists respectively.Third,PrePost can directly find frequent itemsets without generating candidate itemsets in some cases by making use of the single path property of N-list.We have experimentally evaluated PrePost against four state-of-the-art algorithms for mining frequent itemsets on a variety of real and synthetic datasets.The experimental results show that the PrePost algorithm is the fastest in most cases.Even though the algorithm consumes more memory when the datasets are sparse,it is still the fastest one.     

完全判别分析

针对现有的线性判别分析算法中存在的降维舍弃空间的判别信息丢失问题,以及秩空间和零空间的判别信息难以兼顾的问题,本文提出了一种完全判别分析算法.算法通过新构建一个子空间以及其中的判别矩阵,实现了可以充分使用全空间的判别信息;且新空间的维数较低,算法流程简单,计算代价较小.相关实验结果证实了本文算法较传统判别分析算法有更好的性能和效率.     

基于学习的高分辨率掌纹细节点质量评价方法

细节点在高分辨率掌纹匹配中扮演了重要角色,然而掌纹图像受到主线、褶皱线等的影响,提取出的细节点质量参差不齐.所以,对细节点进行质量评价并去除伪细节点,成为一个研究课题.提出了一种基于学习的高分辨率掌纹细节点质量评价方法.首先使用了基于图像的Gabor卷积响应和复数滤波响应等的一系列特征,用来对细节点局部进行冗余描述;然后,把每个特征作为弱分类器,用AdaBoost算法进行多层训练,挑选出对真伪细节点判别效果最理想的特征;最后,把弱分类器加权线性组合的响应分数作为细节点质量的得分,筛选出得分在阈值以上的细节点作为真细节点.该方法的实验结果与基于傅里叶变换的方法相比,能够更好地区分真伪细节点,对细节点的质量做出了更好的评价.     

SPG-Suite：面向伪周期时间序列的预测方法

伪周期时间序列是一种广泛存在的数据形式,它具有伪周期性、非平稳性和特征值等特征。对这类时间序列进行预测,具有很强的研究和应用意义。然而,目前的相关研究对伪周期时间序列的关注度不足,一些已有的时间序列预测方法在应用到伪周期时间序列上时,会造成误差的累积,使得预测效果很差。为了解决这些问题,总结了伪周期时间序列的特征,并提出了SPG-Suite预测方法,很好地解决了传统方法无法解决的问题。最后,在真实的数据集上进行了实验,并与传统方法进行了对比,实验结果表明,SPG-Suite方法在预测精度上具有明显的优势,并具有较强的可扩展性。     

NeurIPS 2020观察与分析

神经信息处理系统大会（Conference on Neural Information Processing Systems,NeurIPS）是机器学习领域的顶级会议,在中国计算机学会（China Computer Federation,CCF）推荐国际学术会议中被评为人工智能领域的A类会议,一直广受关注。NeurIPS 2020收到了创纪录的9 467篇投稿,最终录用1 898篇论文。收录的论文涵盖了人工智能的各种主题,包括深度学习及其应用、强化学习与规划、纯理论研究、概率方法、优化及机器学习与社会等。本文回顾了NeurIPS 2020的亮点及论文录用情况,详细解读了特邀报告、最佳论文、口头报告及部分海报论文,希望能帮助读者快速了解NeurIPS 2020的盛况。     

Controllability of multi-agent systems based on agreement protocols

This paper investigates the controllability of multi-agent systems based on agreement protocols. First, for a group of single-integrator agents, the controllability is studied in a unified framework for both networks with leader-following structure and networks with undirected graph. Some new necessary/sufficient conditions for the controllability of networks of single-integrator agents are established. Second, we prove that, under the same topology and same prescribed leaders, a network of high-order dynamic agents is completely controllable if and only if so is a network of single-integrator agents. Third, how the selection of leaders and the coupling weights of graphs affect the controllability is analyzed. Finally, some numerical simulations are presented to demonstrate the effectiveness of the theoretical results.     

Collective motion in non-reciprocal swarms

This paper studies a non-reciprocal swarm model that consists of a group of mobile autonomous agents with an attraction-repulsion function governing the interaction of the agents. The function is chosen to have infinitely large values of repulsion for vanishing distance between two agents so as to avoid occurrence of collision. It is shown analytically that under the detailed balance condition in coupling weights, all the agents will aggregate and eventually form a cohesive cluster of finite size around the weighted center of the swarm in a finite time. Moreover, the swarm system is completely stable, namely, the motion of all agents converge to the set of equilibrium points. For the general case of non-reciprocal swarms without the detailed balance condition, numerical simulations show that more complex self-organized oscillations can emerge in the swarms. The effect of noise on collective dynamics of the swarm is also examined with a white Gaussian noise model.     

基于带有惩罚因子的阴性选择算法的恶意程序检测模型

提出了一个基于带有惩罚因子的阴性选择算法的恶意程序检测模型.该模型从指令频率和包含相应指令的文件频率两个角度出发,对指令进行了深入的趋向性分析,提取出了趋向于代表恶意程序的恶意程序指令库.利用这些指令,有序切分程序比特串,模型提取得到恶意程序候选特征库和合法程序类恶意程序特征库.在此基础上,文中提出了一种带有惩罚因子的阴性选择算法(negative selection algorithm with penalty factor,NSAPF),根据异体和自体的匹配情况,采用惩罚的方式,对恶意程序候选特征进行划分,组成了恶意程序检测特征库1(malware detection signature library 1,MDSL1)和恶意程序检测特征库2(MDSL2),以此作为检测可疑程序的二维参照物.综合可疑程序和MDSL1,MDSL2的匹配值,文中模型将可疑程序分类到合法程序和恶意程序.通过在阴性选择算法中引入惩罚因子C,摆脱了传统阴性选择算法中对自体和异体有害性定义的缺陷,继而关注程序代码本身的危险性,充分挖掘和调节了特征的表征性,既提高了模型的检测效果,又使模型可以满足用户对识别率和虚警率的不同要求.综合实验...