Partial Sum Minimization of Singular Values in Robust PCA: Algorithm and Applications

Tae-Hyun Oh,Yu-Wing Tai,Bazin, Jean-Charles,Hyeongwoo Kim,In So Kweon IEEE 2016 IEEE transactions on pattern analysis and machine Vol.38 No.4

<P>Robust Principal Component Analysis (RPCA) via rank minimization is a powerful tool for recovering underlying low-rank structure of clean data corrupted with sparse noise/outliers. In many low-level vision problems, not only it is known that the underlying structure of clean data is low-rank, but the exact rank of clean data is also known. Yet, when applying conventional rank minimization for those problems, the objective function is formulated in a way that does not fully utilize a priori target rank information about the problems. This observation motivates us to investigate whether there is a better alternative solution when using rank minimization. In this paper, instead of minimizing the nuclear norm, we propose to minimize the partial sum of singular values, which implicitly encourages the target rank constraint. Our experimental analyses show that, when the number of samples is deficient, our approach leads to a higher success rate than conventional rank minimization, while the solutions obtained by the two approaches are almost identical when the number of samples is more than sufficient. We apply our approach to various low-level vision problems, e.g., high dynamic range imaging, motion edge detection, photometric stereo, image alignment and recovery, and show that our results outperform those obtained by the conventional nuclear norm rank minimization method.</P>

Model-Based Test-Suite Minimization using Modified Condition/Decision Coverage (MC/DC)

Ajay Kumar Jena,Santosh Kumar Swain,Durga Prasad Mohapatra 보안공학연구지원센터 2015 International Journal of Software Engineering and Vol.9 No.5

Testing is very expensive for high-assurance software, like commercial aircraft systems, weapon research, weather forecast, earthquake forecast, and software used for safety critical system. A small and simple flaw in the end product can be enough for destroying the entire effort of the developer with a huge unrecoverable damage to the society. For this reason, Federal Aviation Administration’s requirement is that, the test-suites should be comprises of Modified Condition/Decision Coverage (MC/DC) adequate. By using logic coverage criteria lots of flaws can be removed for safety critical software. MC/DC was proposed by NASA, and had been widely accepted in the field of testing. MC/DC is an effective verification technique, and helps to uncover safety faults. It is a challenge to minimize the number of test-suites when there is a partial change in the software. This can be achieved by using models. Unified Modeling Language (UML) not only helps to design software but also plays a vital role in detecting the faults early phase of design and in minimizing the test-suite. Existing test-suite minimization techniques investigated by different researchers may not be effective in minimizing MC/DC-adequate test-suites because they do not consider the complexity of the present software. A new approach for test-suite minimization is presented in this work, using dissimilarity matrix, which can be well fitted with MC/DC. We also present the results generated out of a case study of the test-suite minimization.

지속적인 세무전략이 기업신용등급에 미치는 영향

신상이,박종일 한국세무사회 부설 한국조세연구소 2019 세무와 회계 연구 Vol.8 No.2

We investigate whether firms with sustainable tax strategies benefit from higher credit ratings relative to firms implementing minimization tax strategies. Recent empirical and anecdotal evidence suggests that firms’ tax strategies have at least two recognizable dimensions, sustainability and minimization(e.g., McGuire et al. 2013;Neuman et al. 2016 etc.). Tax strategies focusing on sustainability strive to achieve a consistent reported tax outcome over time [i.e., a narrow range of cash effective tax rates(cash ETRs)], with less emphasis on the outcome’s level. In contrast, tax strategies that focus on minimization(i.e., the level of tax avoidance) seek to reduce firms’ explicit tax burdens and achieve the lowest possible tax outcome(i.e., a low ETR). However, studies that focus on the level of a firm’s tax avoidance implicitly assume that the primary objective a firm’s tax strategy is to minimize taxes(Shackelford and Shevlin 2001;Scholes et al. 2008). Scholes et al.(2008) develop a framework that suggests that the ultimate goal of a firm’s tax strategy is to maximize after-tax returns, which is not necessarily achieved by minimizing taxes, noting that effective tax planning is not equivalent to tax minimization. For examples, prior researches find that firms with sustainable tax strategies have more persistent pre-tax earnings and earnings components and also find that investors are able to infer the sustainability of a firm’s tax strategy and use it as a signal to assess the persistence of pre-tax earnings and earnings components(McGuire et al. 2013). Neuman et al.(2013) find that firms with sustainable tax strategies have more transparent information environments and Neuman(2016) find that firms implementing sustainable tax strategies earn significantly higher median buy-and-hold returns than firms implementing minimization tax strategies. Therefore, we extend the prior research by providing novel evidence on credit rating agency use it as a positive signal to assess the sustainable of a firm’s tax strategy. For this analysis, following the methodology of prior research that we employ empirical measures of the sustainability of firm’s tax strategies Specifically, we measure the coefficient of variation for annual cash ETRs for the five-year period that includes the current year’s cash ETR and the four prior years’ cash ETRs (i.e., over the period t-4 to t). The coefficient of variation for cash ETRs (hereafter SUSTAIN) is the standard deviation of annual cash ETRs(hereafter TR) scaled by the absolute value of the mean of annual cash ETRs over the same five-year period(e.g., McGuire et al. 2013;Neuman 2016) and we measure aggressive tax avoidance as if its five-year cash ETR, measured from year t-4 to t(Dyreng et al. 2008), is in the lowest quintile(Donohoe and Knechel 2014). Following the prior studies(Park and Yoon 2013;Jo 2016), this paper employ corporate credit rating come from KISVALUE. Our sample of observations covers KOSPI and KOSDAQ listed firms based on the test variable(based on the dependent variable) from 2003 to 2016(from 2004 to 2017), thus we use 8,328 firm-year observations. Our empirical results reveal the following. After controlling for several factors that affect credit rating, we find that positive and significant coefficients on SUSTAIN(TR) in all specifications of the model(all p < 0.01), which suggests that, at the average, firms with sustainable tax strategies have significantly higher credit rating. Whereas, we find that negative and significant coefficients on tax avoidance, firms with minimization tax strategies have significantly lower credit rating, which suggesting that minimization may not be the only method of tax planning that maximizes value. Our overall inferences are unchanged when using this alternative measure of cash ETR and GAAP ETR. In addition, when we divided the sample into KOSPI and KOSDAQ listed firms, we find that obtain similar res... 본 연구는 지속적인 세무전략이 신용평가기관의 신용등급에 긍정적인 영향을 주는지를 실증적으로 분석하였다. 또한 앞서의 주제를 분석할 때 법인세최소화를 위한 공격적 조세회피의 세무전략과도 비교하여 살펴보았다. 선행연구에서는 현금유효세율의 변동성을 낮게 유지하는 기업의 세무전략이 이익지속성이 높고, 더 투명한 정보환경에 있으며, 투자자 측면에서 가격결정에 반영된다는 결과를 보고하였다(McGuire et al. 2013;Neuman et al. 2013;Neuman 2016). 따라서 본 연구는 앞서의 선행연구의 범위를 보다 확장시켜 신용평가기관 측면에서 지속적인 세무전략의 정보를 기업신용등급에 긍정적으로 반영하는지를 살펴보고자 한다. 분석을 위해 지속적인 세무전략의 측정은 앞서의 선행연구의 방법으로 측정되었고, 공격적 조세회피는 Donohoe and Knechel(2014)의 방법에 따라 지시변수로 측정하였다. 분석기간은 관심변수를 기준으로 2003년부터 2016년까지이고, 최종표본은 12월 결산의 상장기업을 대상으로 8,328개 기업/연 자료이다. 실증분석 결과, 신용등급에 영향을 주는 통제변수가 고려된 후에도 지속적인 세무전략의 기업일수록 기업신용등급에 긍정적인 영향을 주는 것으로 나타났다. 이와 달리, 법인세최소화를 위한 공격적 조세회피의 세무전략은 기업신용등급에 부정적인 영향을 주는 것으로 나타났다. 이러한 결과는 Cash ETR과 GAAP ETR로 측정한 경우 모두 일치된 결과를 보였다. 또한 전체표본을 다시 KOSPI와 KOSDAQ 표본으로 나누어 분석하더라도 모두 앞서와 일치된 결과로 나타났다. 이상의 본 연구결과는 현금유효세율의 변동성을 낮게 유지하는 기업의 지속적인 세무전략이 신용평가기관에게 신용위험을 낮추는 신호를 제공하는 데 반해, 법인세최소화를 위한 조세회피의 세무전략은 신용위험을 높이는 정보를 제공한다는 것을 비교하여 보여주었다는 데 의미가 있다. 또한 지속적인 세무전략과 신용등급 간의 관계를 처음으로 살펴본 연구라는 점에서 의의가 있다. 본 연구의 발견은 그동안 법인세최소화가 기업가치에 긍정적일 것으로 보는 관점과 달리, 선택적인 접근으로 유효세율의 변동성을 낮게 유지하는 기업의 세무전략이 신용등급 측면에서 더 효과적인 세무계획일 수 있음을 시사해 준다. 이와 더불어 조세회피를 위한 세무전략이 신용평가 측면에서 부정적인 효과가 나타난 결과 역시 Shackelford and Shevlin (2001)의 효과적인 세무계획 관점에서 보면 비세금비용을 증가시킨다는 점에서 기업가치의 극대화를 위한 유일한 세무전략의 방법이 아닐 수 있음을 시사한다. 한편으로, 본 연구결과는 지속적인 세무전략의 정보가 신용등급의 결정에 중요한 변수임을 시사한다.

Loss Minimization Control of Interior Permanent Magnet Synchronous Motors Considering Self-Saturation and Cross-Saturation

Hamidreza Pairo,Mohammad Khanzade,Abbas Shoulaie 전력전자학회 2018 JOURNAL OF POWER ELECTRONICS Vol.18 No.4

In this paper, a loss minimization control method for interior permanent magnet synchronous motors is presented with considering self-saturation and cross saturation. According to variation of the d-axis and q-axis inductances by different values of the d-axis and q-axis components of currents, it is necessary to consider self-saturation and cross saturation in the loss minimization control method. In addition, the iron loss resistance variation due to frequency variation is considered in the condition of loss minimization. Furthermore, the loss minimization control method is compared with maximum torque per ampere (MTPA), unity power factor (UPF) and id=0 control methods. Experimental results verify the performance and proper dynamic response of the loss minimization control method with considering self-saturation and cross saturation.

Loss Minimization Control of Interior Permanent Magnet Synchronous Motors Considering Self-Saturation and Cross-Saturation

Pairo, Hamidreza,Khanzade, Mohammad,Shoulaie, Abbas The Korean Institute of Power Electronics 2018 JOURNAL OF POWER ELECTRONICS Vol.18 No.4

In this paper, a loss minimization control method for interior permanent magnet synchronous motors is presented with considering self-saturation and cross saturation. According to variation of the d-axis and q-axis inductances by different values of the d-axis and q-axis components of currents, it is necessary to consider self-saturation and cross saturation in the loss minimization control method. In addition, the iron loss resistance variation due to frequency variation is considered in the condition of loss minimization. Furthermore, the loss minimization control method is compared with maximum torque per ampere (MTPA), unity power factor (UPF) and $i_d=0$ control methods. Experimental results verify the performance and proper dynamic response of the loss minimization control method with considering self-saturation and cross saturation.

REGULATORY FOCUS AND DECISION RULES: ARE PREVENTIONFOCUSED CONSUMERS REGRET MINIMIZERS?

Jooyoung Lim,Minhi Hahn 글로벌지식마케팅경영학회 2018 Global Marketing Conference Vol.2018 No.07

When choosing, people often anticipate how they might feel about future outcomes, and use those feelings as guides to choices. Although the impact of emotions on decision-making is well-documented, both theoretically and empirically, relatively limited efforts have been made to quantitatively incorporate such anticipated emotions, specifically regret, into the discrete choice model, which is a workhorse for investigating various consumer choices in marketing. Regret theory suggests that the most relevant emotion in the context of decision-making is regret, which is the emotion that we experience when realizing or imagining that our current situation would have been better, if only we had decided differently. Consumers are known to be regret-averse and motivated to avoid regret. What is unique about regret is that it is directly linked to the choice or decision at hand, while other emotions relevant to decision-making may occur in the absence of a decision because these emotions are related to aspects of outcomes or uncertainty. In transportation science, a new econometric, multinomial, multi-attribute reference-dependent discrete choice model of random regret minimization (RRM) based on regret theory has recently been proposed. In this paradigm, individuals are assumed to make decisions so as to minimize the anticipated regret, which is experienced when a foregone alternative performs better than the chosen alternative at attribute-levels. However, there have been limited efforts to utilize the RRM to investigate consumers’ choices although marketing has long been interested in understanding consumers’ choices. Furthermore, despite the growing interest in the RRM in other disciplines, relatively little is known about the potential drivers for the decision rules of utility-maximization and regret-minimization. In this research, we attempt to shed light on a potential consumer-specific driver for decision rules between random utility maximization and random regret minimization in discrete choice modeling to gain a better understanding of consumers’ decision process, while introducing the RRM in the marketing domain. Specifically, we investigate what types of consumers are more likely to be regret-minimizers or utility-maximizers based on regulatory focus theory. We posit that chronically prevention-focused consumers are more likely to be regret-minimizers while chronically promotion-focused consumers are more likely to be utility-maximizers. We employ a latent class modeling framework to incorporate structural heterogeneity of decision rules to test the hypotheses using a discrete choice experiment on US residents. Notably, while the hypotheses are supported, a portion of prevention-focused consumers are in fact more likely to be utility-maximizers rather than regret-minimizers, indicating that the consumers’ regulatory focus is not entirely mapped with the decision rules on a one-to-one basis. The empirical finding further suggests that consumer-specific variables other than consumers’ chronic regulatory focus may be useful for identifying a regret-minimizing segment.

지속적인 세무전략이 정보비대칭과 자본비용에 미치는 영향

박종일 한국세무사회 부설 한국조세연구소 2019 세무와 회계 연구 Vol.8 No.3

This study investigate whether firms with sustainable tax strategies is associated with lower information asymmetry and lower cost of capital. For the purpose of comparison, I also investigate whether the firms with minimization tax strategies is associated with information asymmetry and higher cost of capital. Prior research indicates firms with sustainability tax strategies is the “dimension of a firm’s tax strategy that focuses on maintaining consistent tax outcomes over time [i.e., a narrow range of effective tax rates (ETRs)]” (McGuire et al. 2013). Prior research documents that firms with sustainable tax strategies experience more persistent earnings (McGuire et al. 2013). Neuman et al. (2013) shows that firms with sustainable tax strategies are associated with more transparent information environments. Therefore, I extend this line of research expect firms with sustainable tax strategies to exhibit lower information asymmetry and lower cost of capital. Scholes et al. (2008) suggest that such myopic tax minimization is not equivalent to effective tax planning;rather, effective tax planning incorporates non-tax strategies and costs to improve overall after-tax returns and firm value. For this analysis, I construct a measure of tax strategy sustainability (hereafter STS) using the coefficient of variation for annual Cash (GAAP) ETRs over a five-year period following McGuire et al. (2013). This is, the coefficient of variation for Cash (GAAP) ETRs is the standard deviation of annual Cash (GAAP) ETRs scaled by the absolute value of mean of Cash (GAAP) ETRs over the same five-year period (i.e., over the period t-4 to t). And I employ tax avoidance (hereafter TAVO) as if its five-year Cash (GAAP) ETR in the long-terms, measured from year t-4 to t. I exploit information asymmetry (hereafter VOLA) is calculated by standard deviation of daily stock returns and cost of capital (hereafter COC) employ WACC comes from KISVALUE databases. I test related each STS (or TAVO) measure to one-year-ahead VOLA and COC. This study sample covers KOSPI and KOSDAQ listed firms based on the test variable (dependent variable) from 2003 to 2017 (from 2004 to 2018), final samples is 9,146 firm-year observations. This study empirical results reveal the following. First, after controlling for several variables that affect the dependent variable (VOLA), I expect to find that a negative and significant coefficients on STS in all specifications of the model, which suggests that firms with sustainable tax strategies have lower information asymmetry. In contrast, I find that positive and significant coefficients on TAVO, which indicating that firms with minimization tax strategies have significantly higher information asymmetry. Second, after controlling for several variables that affect the dependent variable (COC), I expect to find that a negative and significant coefficients on STS in all specifications of the model, which suggests that firms with sustainable tax strategies have lower cost of capital. In contrast, I find that positive and significant coefficients on TAVO, which indicating that firms with minimization tax strategies have significantly higher cost of capital. These results are qualitatively similar when using a alternative measure as GAAP ETR instead of Cash ETR. Furthermore, my results are robust to controlling for pre-tax earnings volatility. In sum, I find that the sustainability of a firm’s strategy is negatively associated with information asymmetry and cost of capital, which is consistent with Scholes et al.’s (2008) argument that effective tax planning may not be focused solely on tax minimization. Thus, my results is consistent with the Scholes et al.’s framework. This study results suggest that the most effective tax strategy for a firm is focused on sustainable tax strategies rather than tax minimization. This is the first study to address the association between sustainable tax strategies and information... 본 논문은 지속적인 세무전략이 정보비대칭 및 자본비용에 미치는 영향을 실증적으로 규명하는 데 있다. 또한 앞서의 관계를 알아보는 데 있어 법인세최소화를 위한 조세회피 세무전략과 비교하여 살펴보았다. 선행연구들은 현금유효세율의 변동성을 낮게 유지하는 기업의 지속적인 세무전략이 이익의 지속성이 높고(McGuire et al. 2013), 또한 더 투명한 정보환경에 있음을 보고하였다(Neuman et al. 2013). 따라서 본 연구는 앞서의 선행연구의 범위를 확장시켜 지속적인 세무전략과 차기의 정보비대칭 및 자본비용 간에 음(-)의 관련성이 있는지를 유가증권과 코스닥기업을 대상으로 알아보았다. 분석을 위해 지속적인 세무전략은 McGuire et al.(2013)의 연구에 따라 유효세율의 변이계수로 측정하였고, 조세회피는 과거 5년간(t-4∼t) Cash ETR과 GAAP ETR의 장기유효세율을 이용하였다. 관심변수(종속변수)를 기준으로 분석기간은 2003년부터 2017년까지(2004년부터 2018년까지)이고, 최종표본은 12월 결산의 9,146개 기업/연 자료이다. 실증분석 결과는 지속적인 세무전략과 차기의 정보비대칭(1년간의 일별주식수익률의 표준편차)은 1% 이내에서 유의한 음(-)의 관계를, 또한 지속적인 세무전략과 차기의 가중평균자본비용은 1% 이내에서 유의한 음(-)의 관계를 보였다. 반면에, 법인세최소화를 위한 조세회피 세무전략은 차기의 정보비대칭 또는 가중평균자본비용과 각각 유의한 양(+)의 관계였다. 이러한 결과는 Cash ETR 대신 GAAP ETR로 분석한 경우도 일치하였다. 이상의 본 연구결과는 유효세율의 변동성을 낮게 유지하는 지속적인 세무전략의 기업일수록 정보비대칭이 더 낮고, 자본비용 역시 낮은 반면에, 법인세최소화를 위해 유효세율 수준을 낮추는 조세회피 세무전략의 기업일수록 정보비대칭이 높고, 자본비용 역시 높음을 비교해서 보여주었다는 데 의미가 있다. 특히 본 연구는 지속적인 세무전략과 정보비대칭 또는 자본비용 간의 관계를 처음으로 살펴본 연구라는 점에서 의의가 있다. 나아가 앞서와 같은 본 연구결과는 Scholes et al.(2008)에서 효과적인 세무전략이 기업가치를 극대화하는 데 있다는 주장에 기초해 볼 때 지속적인 세무전략이 조세회피 세무전략보다 더 효과적인 세무전략 중 하나일 수 있음을 비교하여 보여주었다는 점도 관련연구에 추가적인 공헌과 시사점을 제공한다.

비 반복적 에너지 최소화를 통한 빠른 움직임 예측 알고리즘

최두섭(Dooseop Choi),최혁(Hyuk Choi),송원석(Wonseok Song),김태정(Taejeong Kim) 한국정보과학회 2013 정보과학회논문지 : 소프트웨어 및 응용 Vol.40 No.12

본 논문은 에너지 최소화 기법에 기반 하는 빠른 움직임 벡터장 예측 방법을 제안한다. 비디오 프레임의 실제 움직임 벡터장은 제안하는 에너지 함수를 최소화함으로써 예측되며 이 때 에너지 함수는 블록 정합 오류와 움직임 벡터장 내 벡터들의 유사도를 평가하는 smoothness constraint로 구성된다. 제안하는 smoothness constraint는, 움직임 벡터장의 움직임 경계를 보존하기 위하여, 가운데 블록의 블록정합 오류를 최소화 하는 이웃 움직임 벡터가 가운데 블록의 움직임 벡터 예측에 많은 영향을 미치도록 설계되었다. 블록 정합 오류와 smoothness constraint의 균형을 유지하는 가중치 인자는 목표 블록의 이미지 특징을 반영하도록 설계되어 질감이 매우 적거나 혹은 질감이 반복되는 영역에 있는 블록의 실제 움직임 벡터를 효율적으로 예측할 수 있도록 하였다. 한편 에너지 함수의 최소화는 계산량을 획기적으로 줄이기 위하여 인과 문제의 해결 및 vector median filter의 이용을 통하여 비 반복적으로 이루어진다. 실험 결과는 제안하는 알고리즘이 더욱 적은 계산량으로 최신 움직임 예측 알고리즘과 비교할만한 성능을 나타냄을 보여주고 있다. In this paper, we propose a fast motion estimation algorithm based on energy minimization framework. The true motion vector field (MVF) of a video frame is estimated by minimizing the energy function, which is defined as the weighted sum of the block-matching errors and the irregularity measure of the motion vector field (MVF), called smoothness constraint. The smoothness constraint proposed in this paper is designed such that the motion vector (MV) of the center block is mainly affected by the neighboring MVs that minimize block-matching errors of the center block in order to preserve motion boundaries. The weight factor, which trades off the block-matching errors and the smoothness constraint, is adaptively determined depending on the image characteristic of the target block so the true MV of the block with low textures or in periodic patterned areas is correctly estimated through the energy minimization. The minimization of the energy function, on the other hand, is done in a non-iterative manner by solving causality problem and exploiting vector median filter in order to drastically reduce computational complexity. Experimental results show that the proposed algorithm shows comparable performance to the state-of-the-art motion estimation algorithms at lower computational complexity.

Intensity-guided edge-preserving depth upsampling through weighted L₀ gradient minimization

Jung, Cheolkon,Yu, Shengtao,Kim, Joongkyu Elsevier 2017 Journal of visual communication and image represen Vol.42 No.-

<P><B>Abstract</B></P> <P>Depth is an important visual cue to perceive real-world scenes. Although a time-of-flight (ToF) depth camera can provide depth information in dynamic scenes, captured depth images are often noisy and of low resolution. In this paper, we propose an intensity-guided edge-preserving depth upsampling method through weighted L<SUB>0</SUB> gradient minimization to enhance both resolution and visual quality of depth images. Guided by the high-resolution intensity image, we perform optimization to preserve boundaries of objects. We apply L<SUB>0</SUB> gradient to the regularization term, and compute its weight from both intensity and depth images. We optimize the objective function using alternating minimization and half-quadratic splitting. Experimental results on Middlebury 2005, 2014, and real-world scene datasets demonstrate that the proposed method produces boundary-preserving depth upsampling results and outperforms state-of-the-art ones in terms of accuracy.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Intensity-guided edge-preserving depth upsampling. </LI> <LI> Weighted L<SUB>0</SUB> gradient minimization. </LI> <LI> Alternating minimization and half-quadratic splitting. </LI> <LI> Suppressing edge blurring and texture copying artifacts. </LI> </UL> </P> <P><B>Graphical abstract</B></P> <P>[DISPLAY OMISSION]</P>

Determination of the sample size and the cost minimization: Progressive stress accelerated degradation test based on the Wiener process

Siil Sung 한국신뢰성학회 2019 International Journal of Reliability and Applicati Vol.20 No.2

Sung (2019) develops the degradation path model for the optimal accelerated degradation test plan under the progressive-stress loading based on the Wiener process. In this paper, the result from Sung (2019)’s research is utilized for developing the problem of the determination of the sample size and the cost minimization. In addition, the case study and sensitivity analysis results are presented based on the proposed method for utilizing for the real situation.