Robust principal component analysis? Candès, Emmanuel J.; Li, Xiaodong; Ma, Yi ...
Journal of the ACM,
05/2011, Letnik:
58, Številka:
3
Journal Article
Recenzirano
Odprti dostop
This article is about a curious phenomenon. Suppose we have a data matrix, which is the superposition of a low-rank component and a sparse component. Can we recover each component individually? We ...prove that under some suitable assumptions, it is possible to recover both the low-rank and the sparse components
exactly
by solving a very convenient convex program called
Principal Component Pursuit
; among all feasible decompositions, simply minimize a weighted combination of the nuclear norm and of the ℓ
1
norm. This suggests the possibility of a principled approach to robust principal component analysis since our methodology and results assert that one can recover the principal components of a data matrix even though a positive fraction of its entries are arbitrarily corrupted. This extends to the situation where a fraction of the entries are missing as well. We discuss an algorithm for solving this optimization problem, and present applications in the area of video surveillance, where our methodology allows for the detection of objects in a cluttered background, and in the area of face recognition, where it offers a principled way of removing shadows and specularities in images of faces.
Long-term operational stability is the foremost issue delaying the commercialization of perovskite solar cells (PSCs). Here we demonstrate an in-situ cross-linking strategy for operationally stable ...inverted MAPbI
PSCs through the incorporation of a cross-linkable organic small molecule additive trimethylolpropane triacrylate (TMTA) into perovskite films. TMTA can chemically anchor to grain boundaries and then in-situ cross-link to a robust continuous network polymer after thermal treatment, thus enhancing the thermal, water-resisting and light-resisting properties of organic/perovskite films. As a result, the cross-linked PSCs exhibit 590-fold improvement in operational stability, retaining nearly 80% of their initial efficiency after continuous power output for 400 h at maximum power point under full-sun AM 1.5 G illumination of Xenon lamp without any UV-filter. In addition, under moisture or thermal (85 °C) conditions, cross-linked TMTA-based PSCs also show excellent stability with over 90% of their initial or post burn-in efficiency after aging for over 1000 h.
The circadian clock regulates daily changes in behavioral, endocrine, and metabolic activities in mammals. Circadian rhythms in cellular physiology are significantly affected by aging. In particular, ...we previously found that aging has a profound impact on daily rhythms in mitochondrial functions in mouse liver, leading to increased oxidative stress. This is not due to molecular clock malfunctions in peripheral tissues in old mice, however, as robust clock oscillations are observed therein. Nonetheless, aging induces changes in gene expression levels and rhythms in peripheral and probably central tissues. In this article, we review recent findings on the roles of the circadian clock and the aging process in regulating mitochondrial rhythms and redox homeostasis. Chronic sterile inflammation is implicated in mitochondrial dysfunction and increased oxidative stress during aging. In particular, upregulation of the NADase CD38 by inflammation during aging contributes to mitochondrial dysregulation.
In this paper we consider a system of quadratic equations $|\langle \bm{z_j}, \bm{x}\rangle|^2=b_j, j=1,\ldots,m$, where $\bm{x} \in \mathbb{R}^n$ is unknown while normal random vectors $\bm{z_j} \in ...\mathbb{R}^n$ and quadratic measurements $b_j \in \mathbb{R}$ are known. The system is assumed to be underdetermined, i.e., $m<n$. We prove that if there exists a sparse solution $\bm{x}$, i.e., at most $k$ components of $\bm{x}$ are nonzero, then by solving a convex optimization program, we can solve for $\bm{x}$ up to a multiplicative constant with high probability, provided that $k\leq O(\sqrt{m\over{\log n}})$. On the other hand, we prove that $k \leq O(\log n\sqrt{m})$ is necessary for a class of natural convex relaxations to be exact. PUBLICATION ABSTRACT
Solar CO2 reduction efficiency is largely limited by poor photoabsorption, sluggish electron–hole separation, and a high CO2 activation barrier. Defect engineering was employed to optimize these ...crucial processes. As a prototype, BiOBr atomic layers were fabricated and abundant oxygen vacancies were deliberately created on their surfaces. X‐ray absorption near‐edge structure and electron paramagnetic resonance spectra confirm the formation of oxygen vacancies. Theoretical calculations reveal the creation of new defect levels resulting from the oxygen vacancies, which extends the photoresponse into the visible‐light region. The charge delocalization around the oxygen vacancies contributes to CO2 conversion into COOH* intermediate, which was confirmed by in situ Fourier‐transform infrared spectroscopy. Surface photovoltage spectra and time‐resolved fluorescence emission decay spectra indicate that the introduced oxygen vacancies promote the separation of carriers. As a result, the oxygen‐deficient BiOBr atomic layers achieve visible‐light‐driven CO2 reduction with a CO formation rate of 87.4 μmol g−1 h−1, which was not only 20 and 24 times higher than that of BiOBr atomic layers and bulk BiOBr, respectively, but also outperformed most previously reported single photocatalysts under comparable conditions.
BiOBr atomic layers with abundant oxygen vacancies were synthesized. The photoresponse of BiOBr extends into the visible range, while charge delocalization around the vacancies contributes to CO2 conversion into COOH*. The material catalyzes visible‐light‐driven CO2 reduction with a CO formation rate of 87.4 μmol g−1 h−1, which is 20 and 24 times greater than that of BiOBr atomic layers and bulk BiOBr, respectively.
The effect of defects on electron–hole separation is not always clear and is sometimes contradictory. Herein, we initially built clear models of two-dimensional atomic layers with tunable defect ...concentrations, and hence directly disclose the defect type and distribution at atomic level. As a prototype, defective one-unit-cell ZnIn2S4 atomic layers are successfully synthesized for the first time. Aberration-corrected scanning transmission electron microscopy directly manifests their distinct zinc vacancy concentrations, confirmed by positron annihilation spectrometry and electron spin resonance analysis. Density-functional calculations reveal that the presence of zinc vacancies ensures higher charge density and efficient carrier transport, verified by ultrafast photogenerated electron transfer time of ∼15 ps from the conduction band of ZnIn2S4 to the trap states. Ultrafast transient absorption spectroscopy manifests the higher zinc vacancy concentration that allows for ∼1.7-fold increase in average recovery lifetime, confirmed by surface photovoltage spectroscopy and PL spectroscopy analysis, which ensures promoted carrier separation rates. As a result, the one-unit-cell ZnIn2S4 layers with rich zinc vacancies exhibit a carbon monoxide formation rate of 33.2 μmol g–1 h–1, roughly 3.6 times higher than that of the one-unit-cell ZnIn2S4 layers with poor zinc vacancies, while the former’s photocatalytic activity shows negligible loss after 24 h photocatalysis. This present work uncovers the role of defects in affecting electron–hole separation at atomic level, opening new opportunities for achieving highly efficient solar CO2 reduction performances.
•Novel feature selection approaches based on Binary Dragonfly Algorithm (BDA) are proposed.•Eight time varying S-shaped and V-shaped transfer functions are proposed.•The leverage of using ...time-varying transfer functions on exploration and exploitation behaviors is investigated.•Extensive tests are made to assess the proposed algorithms on the datasets to prove their merits.
Display omitted
The Dragonfly Algorithm (DA) is a recently proposed heuristic search algorithm that was shown to have excellent performance for numerous optimization problems. In this paper, a wrapper-feature selection algorithm is proposed based on the Binary Dragonfly Algorithm (BDA). The key component of the BDA is the transfer function that maps a continuous search space to a discrete search space. In this study, eight transfer functions, categorized into two families (S-shaped and V-shaped functions) are integrated into the BDA and evaluated using eighteen benchmark datasets obtained from the UCI data repository. The main contribution of this paper is the proposal of time-varying S-shaped and V-shaped transfer functions to leverage the impact of the step vector on balancing exploration and exploitation. During the early stages of the optimization process, the probability of changing the position of an element is high, which facilitates the exploration of new solutions starting from the initial population. On the other hand, the probability of changing the position of an element becomes lower towards the end of the optimization process. This behavior is obtained by considering the current iteration number as a parameter of transfer functions. The performance of the proposed approaches is compared with that of other state-of-art approaches including the DA, binary grey wolf optimizer (bGWO), binary gravitational search algorithm (BGSA), binary bat algorithm (BBA), particle swarm optimization (PSO), and genetic algorithm in terms of classification accuracy, sensitivity, specificity, area under the curve, and number of selected attributes. Results show that the time-varying S-shaped BDA approach outperforms compared approaches.
In particle swarm optimization (PSO), the inertia weight is an important parameter for controlling its search capability. There have been intensive studies of the inertia weight in continuous ...optimization, but little attention has been paid to the binary case. This paper comprehensively investigates the effect of the inertia weight on the performance of binary PSO (BPSO), from both theoretical and empirical perspectives. A mathematical model is proposed to analyze the behavior of BPSO, based on which several lemmas and theorems on the effect of the inertia weight are derived. Our research findings suggest that in the binary case, a smaller inertia weight enhances the exploration capability while a larger inertia weight encourages exploitation. Consequently, this paper proposes a new adaptive inertia weight scheme for BPSO. This scheme allows the search process to start first with exploration and gradually move toward exploitation by linearly increasing the inertia weight. The experimental results on 0/1 knapsack problems show that the BPSO with the new increasing inertia weight scheme performs significantly better than that with the conventional decreasing and constant inertia weight schemes. This paper verifies the efficacy of increasing inertia weight in BPSO.
Ceramic/polymer composite equipped with 3D interlocking skeleton (3D IL) is developed through a simple freeze‐casting method, exhibiting exceptionally light weight, high strength, toughness, and ...shock resistance. Long‐range crack energy dissipation enabled by 3D interlocking structure is considered as the primary reinforcing mechanism for such superior properties. The smart composite design strategy should hold a place in developing future structural engineering materials.
Display omitted
•River δD and δ18O can response to cumulative effects of cascade dams.•Reservoir δD and δ18O are influenced by retention time and mean water depth.•Water δD and δ18O show a jagged ...increase from upstream to downstream of the impounded Wujiang River.
Cascade dams are known to influence the river water cycle, but their cumulative effects (CEs) are still not well understood. Water hydrogen (H) and oxygen (O) isotopes are hypothesized to be used to characterize the CEs. To test this hypothesis, we investigated water δD, δ18O, and related environmental factors in cascade reservoirs on the Wujiang River, Southwest China. The δD and δ18O ranged from −64.2‰ to −45.4‰ and from −9.7‰ to −6.8‰, respectively, and showed obvious temporal and spatial variations. Water temperature is an important factor influencing these variations. After damming, an increase of water retention time caused the enrichment of heavy H-O isotopes in reservoir surface water, and thermal stratification induced a decrease of δD and δ18O with depth. Due to bottom discharge, released water showed more negative δD and δ18O than reservoir surface water, and these δD and δ18O differences were controlled by water retention time and mean water depth of the reservoir. Overall, the CEs of cascade dams caused δD and δ18O to display a jagged increase from upstream to downstream in the impounded Wujiang River. Therefore H-O isotopes can be used to estimate the CEs of cascade dams. As cascade dams can modify H-O isotope signatures, caution should be exercised when using H and O isotopes to trace the source of the impounded river water.