Extraordinarily Time-and Memory-Efficient Large-Scale Canonical Correlation Analysis in Fourier Domain: From Shallow to Deep.

Canonical correlation analysis (CCA) is a correlation analysis technique that is widely used in statistics and the machine-learning community. However, the high complexity involved in the training process lays a heavy burden on the processing units and memory system, making CCA nearly impractical in large-scale data. To overcome this issue, a novel CCA method that tries to carry out analysis on the dataset in the Fourier domain is developed in this article. Appling Fourier transform on the data, we can convert the traditional eigenvector computation of CCA into finding some predefined discriminative Fourier bases that can be learned with only element-wise dot product and sum operations, without complex time-consuming calculations. As the eigenvalues come from the sum of individual sample products, they can be estimated in parallel. Besides, thanks to the data characteristic of pattern repeatability, the eigenvalues can be well estimated with partial samples. Accordingly, a progressive estimate scheme is proposed, in which the eigenvalues are estimated through feeding data batch by batch until the eigenvalues sequence is stable in order. As a result, the proposed method shows its characteristics of extraordinarily fast and memory efficiencies. Furthermore, we extend this idea to the nonlinear kernel and deep models and obtained satisfactory accuracy and extremely fast training time consumption as expected. An extensive discussion on the fast Fourier transform (FFT)-CCA is made in terms of time and memory efficiencies. Experimental results on several large-scale correlation datasets, such as MNIST8M, X-RAY MICROBEAM SPEECH, and Twitter Users Data, demonstrate the superiority of the proposed algorithm over state-of-the-art (SOTA) large-scale CCA methods, as our proposed method achieves almost same accuracy with the training time of our proposed method being 1000 times faster. This makes our proposed models best practice models for dealing with large-scale correlation datasets. The source code is available at https://github.com/Mrxuzhao/FFTCCA.

High performance plant-derived thermoplastic polyester elastomer foams achieved by manipulating charging order of mixed blowing agents.

Plant-derived thermoplastic polyester elastomer (TPEE) is an environment friendly polymer known for its exceptional tear strength and mechanical properties, whose monomers are generated from crops. To prepare high-performance TPEE foams is still challenging due to the intrinsic shrinkage behavior. Herein, two microcellular foaming routes with different charging orders of mixed blowing agents, namely "CO2 firstly charging process (CO2-F-process)" and "N2 firstly charging process (N2-F-process)", were developed to elucidate the effects of mixed blowing agents on foaming behavior. Compared with the case in N2-F-process, more carbon dioxide and less nitrogen were adsorbed in CO2-F-process. Thus, TPEE foams prepared by N2-F-process show less shrinkage and higher creep recovery ratio than those prepared by CO2-F-process. Thanks to better structural stability and smaller shrinkage, TPEE foams prepared by N2-F-process exhibited enhanced strength and resilience. For the foams with similar density, compression strength can be increased by 52 %, and energy loss coefficient can be reduced to 50 %, by using N2-F-process. Thus, not only biomass TPEE foams with enhanced mechanical performance shows promising prospects in those areas that needs lightweight, insulation and high resilience, but also novel microcellular foaming technique with mixed blowing agents opens a new way for developing high-performance polymeric foams.

Absolute measurement approach for crystal growth height based on a polarization-synchronized phase-shifting interferometer.

The quality of the solid deuterium-deuterium (D-D) layer in the inertial confinement fusion (ICF) target plays a vital role in the success of fusion experiments. A good understanding of how the quality is affected by the unstable growth of D-D crystal is required. This article provides an approach of measuring D-D layer absolute height in real time by combining monitoring algorithms and a synchronous phase-shifting interferometer. In the approach taken, a real-time monitoring technology, in which an antivibration algorithm is added, is used to get an absolute height of monitoring zone, overcoming the inability to accurately detect the saltus step in the interferometric measurement. Meanwhile, the polarization-synchronized phase-shifting technology is propitious to retrieve the D-D height distribution in a whole interferogram. Consequently, the categorical altitude of the D-D layer in entire crystalline regions can be obtained. Simulation analysis together with experiments have proved that a non-contact, rapid, and high-precision measurement of the D-D crystal absolute height can be realized by using the interferometer and method proposed.

Tobacco Use and Exposure to Second-Hand Smoke among Urban Residents: A Community-Based Investigation.

OBJECTIVES: In 2005, China acceded to the World Health Organization Framework Convention on Tobacco Control (WHO FCTC), the foundation for the global fight against tobacco. Certain cities in China have established local regulations to control tobacco use ahead of national policy; however, without the enforcement of statutory law, some of these regulations are merely lip service. The aim of the study was to assess the effects of city policy on smoking prevalence and on second-hand smoke (SHS) exposure status among non-smokers in Changchun City. METHODS: A cross-sectional survey covering a multiple-stage, representative sample of the urban population aged ≥15 years was conducted between 1 Dec 2013 and 31 Jan 2014. The WHO and the US Centers for Disease Control and Prevention developed the questionnaires used, which included demographic characteristics, smoking behaviors and SHS exposure status. RESULTS: Overall cigarette smoking prevalence was 23.5%; daily cigarette smoking prevalence was 21.2%. Smoking prevalence and cigarettes consumed per day was higher among men (p < 0.05) and those aged 45-64 years (p < 0.05). Among current smokers, 8.1% planned to quit within 12 months; 53.4% had no intention of quitting. Overall SHS exposure prevalence was 41.9% (workplace) and 34.1% (at home) over the previous 30 days. The weighted workplace SHS exposure prevalence increased with age. CONCLUSIONS: The high proportion of smokers with no intention of quitting and the high level of SHS exposure may constitute one of the most significant barriers to successful smoking cessation in the city. A continued drive to promote full implementation of the WHO FCTC is still needed.