scholarly journals Growing synthetic data through differentially-private vine copulas

2021 ◽  
Vol 2021 (3) ◽  
pp. 122-141
Author(s):  
Sébastien Gambs ◽  
Frédéric Ladouceur ◽  
Antoine Laurent ◽  
Alexandre Roy-Gaumond
Keyword(s):  
Synthesis Method ◽  
Synthetic Data ◽  
Data Synthesis ◽  
Protection Level ◽  
Novel Approach ◽  
Inference Attack ◽  
Arbitrary Dimensions ◽  
Vine Copulas

Abstract In this work, we propose a novel approach for the synthetization of data based on copulas, which are interpretable and robust models, extensively used in the actuarial domain. More precisely, our method COPULA-SHIRLEY is based on the differentially-private training of vine copulas, which are a family of copulas allowing to model and generate data of arbitrary dimensions. The framework of COPULA-SHIRLEY is simple yet flexible, as it can be applied to many types of data while preserving the utility as demonstrated by experiments conducted on real datasets. We also evaluate the protection level of our data synthesis method through a membership inference attack recently proposed in the literature.

Data synthesis method preserving correlation of features

2022 ◽  
Vol 122 ◽  
pp. 108241 ◽  
Author(s):  
Wonseok Yang ◽  
Woochul Nam
Keyword(s):  
Synthesis Method ◽  
Data Synthesis

Synthesis and characterizations of TiN–SBA-15 mesoporous materials for CO2 dry reforming enhancement

2020 ◽  
Vol 92 (4) ◽  
pp. 545-556
Author(s):  
Maslin Chotirach ◽  
Supawan Tantayanon ◽  
Duangamol Nuntasri Tungasmita ◽  
Junliang Sun ◽  
Sukkaneste Tungasmita
Keyword(s):  
Photoelectron Spectroscopy ◽  
Synthesis Method ◽  
High Specific Surface Area ◽  
Dry Reforming ◽  
Dry Reforming Of Methane ◽  
Ni Catalyst ◽  
X Ray ◽  
Novel Approach ◽  
Temperature Programmed ◽  
Adsorption Desorption

AbstractA novel approach of titanium nitride (TiN) incorporated into SBA-15 framework was developed using one-step hydrothermal synthesis method. TiN contents up to ~18 wt% were directly dispersed in a synthetic gel under a typical strong acidic condition. The physico-chemical characteristics and the surface properties were investigated by means of X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Fourier transform infrared spectroscopy (FTIR), N2 adsorption-desorption, field emission scanning electron microscope (FESEM) equipped with energy dispersive X-ray spectroscopy (EDS), wavelength dispersive X-ray fluorescence (WDXRF) and CO2-temperature programmed desorption (CO2-TPD). The results indicated that the highly ordered mesostructured was effectively maintained with high specific surface area of 532–685 m2g−1. The basicity of the modified SBA-15 increased with rising TiN loading. These modified materials were applied as a support of Ni catalyst in dry reforming of methane (DRM). Their catalytic behavior possessed superior conversions for both CO2 and CH4 with the highest H2/CO ratio (0.83) as well as 50 % lower carbon formation, compared to bare SBA-15 support.

scholarly journals Improving Neural Machine Translation by Filtering Synthetic Parallel Data

Entropy ◽  
2019 ◽  
Vol 21 (12) ◽  
pp. 1213
Author(s):  
Guanghao Xu ◽  
Youngjoong Ko ◽  
Jungyun Seo
Keyword(s):  
Machine Translation ◽  
Synthetic Data ◽  
Similarity Score ◽  
Target Language ◽  
Neural Machine Translation ◽  
Novel Approach ◽  
Parallel Data ◽  
Back Translation ◽  
Translation Errors ◽  
Training State

Synthetic data has been shown to be effective in training state-of-the-art neural machine translation (NMT) systems. Because the synthetic data is often generated by back-translating monolingual data from the target language into the source language, it potentially contains a lot of noise—weakly paired sentences or translation errors. In this paper, we propose a novel approach to filter this noise from synthetic data. For each sentence pair of the synthetic data, we compute a semantic similarity score using bilingual word embeddings. By selecting sentence pairs according to these scores, we obtain better synthetic parallel data. Experimental results on the IWSLT 2017 Korean→English translation task show that despite using much less data, our method outperforms the baseline NMT system with back-translation by up to 0.72 and 0.62 Bleu points for tst2016 and tst2017, respectively.

scholarly journals Linear heterostructured Ni2Si/Si nanowires with abrupt interfaces synthesised in solution

Nanoscale ◽  
2018 ◽  
Vol 10 (40) ◽  
pp. 19182-19187 ◽  
Author(s):  
Martin Sheehan ◽  
Quentin M. Ramasse ◽  
Hugh Geaney ◽  
Kevin M. Ryan
Keyword(s):  
Synthesis Method ◽  
Si Nanowires ◽  
Solution Synthesis ◽  
One Pot ◽  
Novel Approach

Herein, we report a novel approach to form axial heterostructure nanowires composed of linearly distinct Ni silicide (Ni2Si) and Si segments via a one-pot solution synthesis method.

Assembly Configuration Synthesis Method for Discretely-Actuated Manipulators

2006 ◽  
Author(s):  
Deanne C. Kemeny ◽  
Raymond J. Cipra
Keyword(s):  
Synthesis Method ◽  
Specific Element ◽  
End Effector ◽  
Novel Approach ◽  
Serial Chain ◽  
Synthesis Strategy ◽  
Planar Grid ◽  
Three Stages ◽  
Configuration Synthesis ◽  
Unique Application

Discretely-actuated manipulators are defined in this paper as serial planar chains of many links where the actuation of one link with respect to the previous link occurs in one of three discrete positions. Because of the limited end-effector workspace, a link may be manually connected to the previous link in one of four 90° orientations to assist in generating a workspace corresponding to specific applications. Given an application workspace, the assembly configuration synthesis strategy presented here is a novel approach to determine the nominal configuration (all actuators in their 0° position) of the serial chain. The solved configuration will cover an application grid area using its discrete actuation with no change in nominal configuration. The unique application workspace, defined as a planar grid area, requires the end effector to be positioned somewhere within each specific element of the grid. The synthesis strategy is made up of three stages with each stage having tests that increase in computation and difficulty that a potential configuration must pass or be eliminated. Critical to the tests is the ability to quickly model and approximate the end-effector workspace of a configuration and a new method for this approximation is described.

Weighted RAIM for APV: The Ideal Protection Level

2010 ◽  
Vol 64 (1) ◽  
pp. 61-73 ◽  
Author(s):  
Carl D. Milner ◽  
Washington Y. Ochieng
Keyword(s):  
Weighted Least Squares ◽  
Failure Detection ◽  
Complex Problem ◽  
International Standards ◽  
Worst Case ◽  
Protection Level ◽  
Novel Approach ◽  
Detection Probabilities ◽  
Receiver Autonomous Integrity Monitoring ◽  
Integrity Risk

International standards require the use of a weighted least-squares approach to onboard Receiver Autonomous Integrity Monitoring (RAIM). However, the protection levels developed to determine if the conditions exist to perform a measurement check (i.e. failure detection) are not specified. Various methods for the computation of protection levels exist. However, they are essentially approximations to the complex problem of computing the worst-case missed detection probability under a weighted system. In this paper, a novel approach to determine this probability at the worst-case measurement bias is presented. The missed detection probabilities are then iteratively solved against the integrity risk requirement in order to derive an optimal protection level for the operation. It is shown that the new method improves availability by more than 30% compared to the baseline weighted RAIM algorithm.A version of this paper was first presented at the US Institute of Navigation (ION) GNSS 2009 Conference in Savannah, Georgia.

scholarly journals Synthesis, Characterizaiton of Cuo Nano-Rod for Supercapacitor Applications

2019 ◽  
Vol 8 (12S) ◽  
pp. 457-458
Keyword(s):  
Electrode Material ◽  
Solvothermal Synthesis ◽  
Synthesis Method ◽  
Electrode System ◽  
Suitable Candidate ◽  
Novel Approach ◽  
Prepared Material ◽  
Supercapacitor Applications ◽  
Cuo Nanorods ◽  
Na2so4 Electrolyte

A novel approach solvothermal synthesis method has been utilized to prepare CuO nanorods for electrochemical capacitors. A new method of synthesis has been adopted for the synthesis of CuO nanostructures. Structural, morphological features of the prepared material were studied by XRD and SEM respectively. Electrochemical supercapacitive performances of the modified electrode material were also analyzed by electrochemical workstation in three-electrode system. This material found to exhibit pesudocapacitive behavior with high capacitance of 135.23 F/g at the prevalent density of 1 A/g in 1M Na2SO4 electrolyte solution, proving a suitable candidate electrode material for supercapacitor applications.

scholarly journals Why Temporal Persistence of Biometric Features, as Assessed by the Intraclass Correlation Coefficient, Is So Valuable for Classification Performance

Sensors ◽  
2020 ◽  
Vol 20 (16) ◽  
pp. 4555
Author(s):  
Lee Friedman ◽  
Hal S. Stern ◽  
Larry R. Price ◽  
Oleg V. Komogortsev
Keyword(s):  
Correlation Coefficient ◽  
Intraclass Correlation Coefficient ◽  
Intraclass Correlation ◽  
Synthetic Data ◽  
Real Data ◽  
Similarity Score ◽  
Classification Performance ◽  
Biometric Analysis ◽  
Novel Approach ◽  
Temporal Persistence

It is generally accepted that relatively more permanent (i.e., more temporally persistent) traits are more valuable for biometric performance than less permanent traits. Although this finding is intuitive, there is no current work identifying exactly where in the biometric analysis temporal persistence makes a difference. In this paper, we answer this question. In a recent report, we introduced the intraclass correlation coefficient (ICC) as an index of temporal persistence for such features. Here, we present a novel approach using synthetic features to study which aspects of a biometric identification study are influenced by the temporal persistence of features. What we show is that using more temporally persistent features produces effects on the similarity score distributions that explain why this quality is so key to biometric performance. The results identified with the synthetic data are largely reinforced by an analysis of two datasets, one based on eye-movements and one based on gait. There was one difference between the synthetic and real data, related to the intercorrelation of features in real data. Removing these intercorrelations for real datasets with a decorrelation step produced results which were very similar to that obtained with synthetic features.

scholarly journals Towards End-to-End Acoustic Localization Using Deep Learning: From Audio Signals to Source Position Coordinates

Sensors ◽  
2018 ◽  
Vol 18 (10) ◽  
pp. 3418 ◽  
Author(s):  
Juan Vera-Diaz ◽  
Daniel Pizarro ◽  
Javier Macias-Guarasa
Keyword(s):  
Three Dimensional ◽  
Synthetic Data ◽  
Audio Signal ◽  
Real Data ◽  
Audio Signals ◽  
Acoustic Source ◽  
Acoustic Localization ◽  
Training Strategy ◽  
Audio Features ◽  
Novel Approach

This paper presents a novel approach for indoor acoustic source localization using microphone arrays, based on a Convolutional Neural Network (CNN). In the proposed solution, the CNN is designed to directly estimate the three-dimensional position of a single acoustic source using the raw audio signal as the input information and avoiding the use of hand-crafted audio features. Given the limited amount of available localization data, we propose, in this paper, a training strategy based on two steps. We first train our network using semi-synthetic data generated from close talk speech recordings. We simulate the time delays and distortion suffered in the signal that propagate from the source to the array of microphones. We then fine tune this network using a small amount of real data. Our experimental results, evaluated on a publicly available dataset recorded in a real room, show that this approach is able to produce networks that significantly improve existing localization methods based on SRP-PHAT strategies and also those presented in very recent proposals based on Convolutional Recurrent Neural Networks (CRNN). In addition, our experiments show that the performance of our CNN method does not show a relevant dependency on the speaker’s gender, nor on the size of the signal window being used.

Sign in / Sign up

Export Citation Format

Share Document