Many big data applications require accurate classification of objects into one possibly thousands or millions of categories. These classification tasks are challenging for various reasons, including class imbalance, high testing cost, and model interpretability problems. To overcome these challenges, we propose a novel hierarchical classification method known as MF-Tree, which stands for matrix factorization tree. Unlike many of the existing methods, our approach is designed to optimize a global objective function. The key theoretical insight of this paper is demonstrating the equivalence between the proposed squared error loss function for matrix factorization and the Hilbert-Schmidt Independence Criterion (HSIC). We showed that the latter has an additive property, thus allowing us to decompose the multi-class learning problem into hierarchical binary classification tasks. To improve its training efficiency, an approximate algorithm for inducing MF-Tree is also proposed. We have performed extensive experiments to compare our methods against several state-of-the-art baseline algorithms. Experimental results suggest that our proposed methods are both effective and efficient when applied to real-world data sets. 

关键词：大型多学习；矩阵分解树；数据分解

查看摘要 索取原文[11页]

Images in reading materials make the content come alive. Aside from providing additional information to the text, reading material containing illustration engages our spatial memory, increases memory retention of the material. However, despite the plethora of available multimedia, adding illustrations to text continues to be a difficult task for the amateur content publisher. To address this problem, we present a semantic-aware image discovery and insertion system for custom publishing. Compared to image search engines, our system has the advantage of being able to discern among different topics within a long text passage and recommend the most relevant images for each detected topic with semantic “visual words” based relevance.

关键词：图像；语义分析；出版定制

查看摘要 索取原文[3页]

We demonstrate post-growth wavelength setting of vertical-cavity surface-emitting lasers (VCSELs) using high-contrast gratings (HCGs). By fabricating HCGs with different duty-cycle and period, the HCG reflection phase can be varied, in effect giving different optical cavity lengths for HCG-VCSELs with different grating parameters. This enables fabrication of monolithic multi-wavelength HCG-VCSEL arrays for wavelength-division multiplexing (WDM). The GaAs HCG is suspended in air by selective removal of an InGaP sacrificial layer. Electrically injected 980-nm HCG-VCSELs with sub-mA threshold currents indicate high reflectivity from the GaAs HCGs. Lasing over a wavelength span of 15 nm was achieved, enabling a 4-channel WDM array with 5 nm channel spacing. Device design, fabrication and experimental proof-of-concept are presented.

关键词：线反射相位；HCG-VCSELs；光栅参数

查看摘要 索取原文[7页]

为了能在云计算环境下安全地使用可搜索加密技术,针对近几年来可搜索加密技术的研究成果,总结了云计算环境下对称/非对称可搜索加密技术的主要算法模型与安全模型,分析了各类安全方案的安全缺陷,并分析了基本可搜索加密方案与加入了密文安全传输、匿名性、第三方代理、隐私保护协议的四类增强方案在基于选择关键字攻击、字典攻击、统计分析攻击、选择明文攻击四类攻击模型下的安全性。最后,总结了可搜索加密技术的应用陷阱,并提出了当前可搜索加密技术面临的安全性挑战。

关键词：可搜索加密；算法模型；安全模型；安全隐患；应用陷阱

查看摘要 索取原文[7页]

基于自主开发的海洋条件系统分析程序RELAP5/MC,研究了倾斜、摇摆海洋条件对浮动式核电厂全船断电事故(SBO)后系统自然循环特性的影响。研究结果表明,对于具有一次侧非能动余热排出系统(PRHR)的浮动式核电厂,PRHR冷凝器在海洋条件下的响应受稳压器波动流量、蒸汽发生器并联热阱的影响,与二次侧PRHR在海洋条件下的性能有较大的差异;稳压器和回路间的流量交换对左舷环路的冷却能力影响较大。

关键词：浮动式核电厂；全船断电；自然循环；RELAP5/MC

查看摘要 索取原文[4页]

This AFOSR-supported research was started in July 2009 and is directed toward the development of a new technology platform for optical and microwave signal processing based on reconfigurable integrated optics devices. This technology has the potential to revolutionize the design circle of optical and high frequency RF systems by providing a common platform for a diverse range of applications. We expect that the impact of the proposed platform in the related fields to be similar to that of field programmable gate arrays (FPGAs) on digital signal processing and digital IC design. The main idea behind RCPA's design is to implement a multi-input multi-output (MIMO) optical processing block that can be adjusted to realize any desired transfer function matrix that lies within the limits defined by the system specifications. The proposed architecture for the RCPA is based on tunable cross-connected element matrices at the input and output and an array of processing unit cells. The proposed architecture will be realized on silicon-on-isolator (SOI) platforms by using coupled micro-resonator and waveguide structures as primary building elements. The basic tunability in the device will be implemented using tunable phase-shifter and couplers based on the thermo-optic effect and free carrier injection. To achieve this goal, in what follows, different steps (including theoretical and modeling tools development, microcavity fabrication and characterization techniques, and methods to develop chip-scale devices) will be presented in detail, with the idea being to address the several challenges before the realization of RCPA chips in both the architecture and device level. Our research in this field has already resulted in a number of scientific publications and technical presentations. A complete list of journal papers and conference presentations is included at the end of this report. AFOSR support has been acknowledged in all these publications and presentations.

关键词：芯片；计算机编程；耦合交互

查看摘要 索取原文[32页]

This report covers direct power injection of 8051-based microcontrollers. A test bench has been designed according to IEC 6213204 standards to test how high frequency signals impact the response of microcontroller circuits in operation. Two microcontroller circuits (Atmel LP2052 and LP216) were tested to the point of upset. Upset is defined as a reduction in voltage or shift in timing from the least significant (LSB) bit output pin (P1.0) of the microcontroller. The chip was exposed to a radio frequency (RF) signal using the direct injection method. A total of five chips were exposed to a continuous wave (CW) signal for up to 1 microsecond. At least 100 shots of pulsed power were applied to the chip from 50 to 200 MHz and each pulsed time based on a single clock cycle. The RF signal applied to the microcontroller in pulse widths of 0.24, 0.48 or 1 microsecond. Each pulse was injected based on target locations or timing positions with respect to the microcontroller clock cycle. Because upsets are different than device failure tests, a probabilistic method was used to establish what power level affected the circuit. This information was converted to probability of effects curves and compared over frequency and the type of board that was used to test the microcontroller.

关键词：芯片(电子)；连续波；无线电信号

查看摘要 索取原文[30页]

The reuse potential of tire chips as coarse aggregates in pavement concrete was examined in this research by investigating the effects of low- and high-volume tire chips on fresh and hardened concrete properties. One concrete control mixture was designed, which well exceeds CDOT Class P concrete requirements. The coarse aggregate component of the mixture was replaced in 100%, 50%, 30%, 20%, and 10% by volume using tire chips. The fresh concrete properties, compressive strength, flexural strength, splitting strength, permeability, and freeze/thaw durability were tested in the lab to evaluate the potential of including tire chips in concrete paving mixes. The testing results indicate tire chips can be used to replace coarse aggregate in concrete pavement mixtures. Two mixtures with 10% coarse aggregate replaced by tire chips had the best performance. The workability was comparable to the control mixture, and the air content reached 6%. At 28 days of age, the average compressive strength of the two mixtures was significantly less than the control but still exceeded CDOTs specification of 4200 psi; the averaged flexural and splitting tensile strengths were higher than 900 psi and 590 psi respectively. In addition, the two mixtures exhibited moderate resistance to chloride-ion penetration at 28 days of age and high freeze/thaw durability. The rubberized mixtures investigated in this study sustained a much higher deformation than the control mixture when subjected to compressive, flexural, and splitting loadings.

关键词：金属芯片；回收材料；比较分析

查看摘要 索取原文[47页]

NASA's Genesis solar wind sample return mission experienced an off nominal landing resulting in broken, albeit useful collectors. Sample 60000 from the collector is comprised of diamond-like-carbon film on a float zone (FZ) silicon wafer substrate Diamond-on-Silicon (DOS), and is highly prized for its higher concentration of solar wind (SW) atoms. A team of scientist at the Johnson Space Center was charged with determining the best, nondestructive and noncontaminating method to subdivide the specimen that would result in a 1 sq. cm subsample for allocation and analysis. Previous work included imaging of the SW side of 60000, identifying the crystallographic orientation of adjacent fragments, and devising an initial cutting plan.

关键词：芯片；成像技术；样本返回任务

查看摘要 索取原文[2页]

Using van der Pauw and microwave surface resistance measurements, a series of temperature-dependent data sets from carbon nanotube (CNT) thin films have been measured. The test structures were fabricated using photolithography, E-beam evaporation, and a novel CNT network deposition technique. The sheet resistance and resistivity of each sample were recorded at temperatures ranging from 0 60 C with test currents ranging from 100 nA to 100 A. These values demonstrated excellent linearity, with no dependence on currents. At room temperature, the sheet resistance yielded a negative temperature coefficient (of approximately 900 ppm/ C). Our objective is to analyze the effect that DC and microwave currents have on CNT thin films using a two-point measurement technique on Corbino discs and in the presence of microwaves (8 12 GHz) to determine at which temperatures a CNT thin film performs best and identify the ideal temperature range in which a CNT microchip yields the maximized sheet resistance. The results counter the trend found by the four-point measurements; a positive temperature coefficient was observed. This observation indicates the temperature coefficient of the material is actually dependent on the size of the sample tested.

关键词：芯片；直流电子束；霍尔效应；温度系数

查看摘要 索取原文[16页]