Scalable Algorithms for Multi-Instance Learning Wei, Xiu-Shen; Wu, Jianxin; Zhou, Zhi-Hua
IEEE transaction on neural networks and learning systems,
04/2017, Volume:
28, Issue:
4
Journal Article
Multi-instance learning (MIL) has been widely applied to diverse applications involving complicated data objects, such as images and genes. However, most existing MIL algorithms can only handle ...small- or moderate-sized data. In order to deal with large-scale MIL problems, we propose MIL based on the vector of locally aggregated descriptors representation (miVLAD) and MIL based on the Fisher vector representation (miFV), two efficient and scalable MIL algorithms. They map the original MIL bags into new vector representations using their corresponding mapping functions. The new feature representations keep essential bag-level information, and at the same time lead to excellent MIL performances even when linear classifiers are used. Thanks to the low computational cost in the mapping step and the scalability of linear classifiers, miVLAD and miFV can handle large-scale MIL data efficiently and effectively. Experiments show that miVLAD and miFV not only achieve comparable accuracy rates with the state-of-the-art MIL algorithms, but also have hundreds of times faster speed. Moreover, we can regard the new miVLAD and miFV representations as multiview data, which improves the accuracy rates in most cases. In addition, our algorithms perform well even when they are used without parameter tuning (i.e., adopting the default parameters), which is convenient for practical MIL applications.
Display omitted
•Mo reduces lattice distortion and increases structural stability of high N MSSs.•Pressurized metallurgy and Mo-alloying increase N solubility in solid solution.•Mo-alloying reduces ...M2N content and alleviates detrimental effect of Cr-depletion.•Mo-alloying significantly improves corrosion resistance of high N MSSs.
In this work, a new high corrosion resistant martensitic stainless steel (MSS) containing high nitrogen content (N > 0.33 wt.%) was designed based on density functional theory (DFT) calculation and pressurized metallurgy. The results indicate that Mo-N alloying significantly improve the corrosion resistance of MSS. This observation has been explained from point of view of crystal structural stability, wherein the reduced formation energy, bonding energy, density of state (DOS) and lattice distortion. It results in the increasing N content in solid solution and the decreasing fractions of Cr depletion induced by the precipitation of Cr2N.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Suppose we are interested in the effect of a treatment in a clinical trial. The efficiency of inference may be limited due to small sample size. However, external control data are often available ...from historical studies. Motivated by an application to Helicobacter pylori infection, we show how to borrow strength from such data to improve efficiency of inference in the clinical trial. Under an exchangeability assumption about the potential outcome mean, we show that the semiparametric efficiency bound for estimating the average treatment effect can be reduced by incorporating both the clinical trial data and external controls. We then derive a doubly robust and locally efficient estimator. The improvement in efficiency is prominent especially when the external control data set has a large sample size and small variability. Our method allows for a relaxed overlap assumption, and we illustrate with the case where the clinical trial only contains a treated group. We also develop doubly robust and locally efficient approaches that extrapolate the causal effect in the clinical trial to the external population and the overall population. Our results also offer a meaningful implication for trial design and data collection. We evaluate the finite‐sample performance of the proposed estimators via simulation. In the Helicobacter pylori infection application, our approach shows that the combination treatment has potential efficacy advantages over the triple therapy.
Full text
Available for:
BFBNIB, DOBA, FSPLJ, FZAB, GIS, IJS, IZUM, KILJ, NLZOH, NUK, OILJ, PILJ, PNG, SAZU, SBCE, SBMB, UILJ, UKNU, UL, UM, UPUK
ABSTRACT
Distributed acoustic sensing provides seismic sensors distributed every few metres over each fibre optic cable of several kilometres in length, serving as an ultra‐dense array for both ...surface and downhole seismic acquisitions. It is well‐recognized now that distributed acoustic sensing acquisition configuration can have profound effect on data quality, notably the choice of the gauge length and lead‐in cable length on the resulting signal to noise ratio. In this paper, we present a detailed study on how these effects behave differently over various signalling frequency bands. We derive the signal to noise ratio as a function of the peak frequency, gauge length and lead‐in cable length, and demonstrate significantly different 80 SNR samples behaviours over low, middle and high frequency bandwidths, respectively. These results, not only reveal the underlying mechanism for signal and noise variation as function of acquisition parameters, but also help identifying the desired gauge and cable lengths that better serve particular applications, such as inversion and imaging which take information from different frequency bands. A field data has been used to demonstrate these frequency‐dependent behaviours.
Full text
Available for:
FZAB, GIS, IJS, KILJ, NLZOH, NUK, OILJ, SBCE, SBMB, UL, UM, UPUK
Current neural networks are mostly built on the MP model, which usually formulates the neuron as executing an activation function on the real-valued weighted aggregation of signals received from ...other neurons. This letter proposes the flexible transmitter (FT) model, a novel bio-plausible neuron model with flexible synaptic plasticity. The FT model employs a pair of parameters to model the neurotransmitters between neurons and puts up a neuron-exclusive variable to record the regulated neurotrophin density. Thus, the FT model can be formulated as a two-variable, two-valued function, taking the commonly used MP neuron model as its particular case. This modeling manner makes the FT model biologically more realistic and capable of handling complicated data, even spatiotemporal data. To exhibit its power and potential, we present the flexible transmitter network (FTNet), which is built on the most common fully connected feedforward architecture taking the FT model as the basic building block. FTNet allows gradient calculation and can be implemented by an improved backpropagation algorithm in the complex-valued domain. Experiments on a broad range of tasks show that FTNet has power and potential in processing spatiotemporal data. This study provides an alternative basic building block in neural networks and exhibits the feasibility of developing artificial neural networks with neuronal plasticity.
Full text
Available for:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
The bioreduction capacity of Cr(VI) by Shewanella is mainly governed by its bidirectional extracellular electron transfer (EET). However, the low bidirectional EET efficiency restricts its wider ...applications in remediation of the environments contaminated by Cr(VI). Cyclic adenosine 3′,5′‐monophosphate (cAMP) commonly exists in Shewanella strains and cAMP–cyclic adenosine 3′,5′‐monophosphate receptor protein (CRP) system regulates multiple bidirectional EET‐related pathways. This inspires us to strengthen the bidirectional EET through elevating the intracellular cAMP level in Shewanella strains. In this study, an exogenous gene encoding adenylate cyclase from the soil bacterium Beggiatoa sp. PS is functionally expressed in Shewanella oneidensis MR‐1 (the strain MR‐1/pbPAC) and a MR‐1 mutant lacking all endogenous adenylate cyclase encoding genes (the strain Δca/pbPAC). The engineered strains exhibit the enhanced bidirectional EET capacities in microbial electrochemical systems compared with their counterparts. Meanwhile, a three times more rapid reduction rate of Cr(VI) is achieved by the strain MR‐1/pbPAC than the control in batch experiments. Furthermore, a higher Cr(VI) reduction efficiency is also achieved by the strain MR‐1/pbPAC in the Cr(VI)‐reducing biocathode experiments. Such a bidirectional enhancement is attributed to the improved production of cAMP–CRP complex, which upregulates the expression levels of the genes encoding the c‐type cytochromes and flavins synthetic pathways. Specially, this strategy could be used as a broad‐spectrum approach for the other Shewanella strains. Our results demonstrate that elevating the intracellular cAMP levels could be an efficient strategy to enhance the bidirectional EET of Shewanella strains and improve their pollutant transformation capacity.
A novel strategy is developed to improve the bidirectional extracellular electron transfer of Shewanella strains in this study. Cheng and coworkers provide the convincing evidences that elevating the intracellular cAMP levels could be used as a simple but efficient strategy to enhance the bidirectional EET of Shewanella strains and improve their pollutant degradation capacity. This strategy could be used as a broad‐spectrum enhancement method for other dissimilatory metal‐reducing bacteria.
Full text
Available for:
BFBNIB, FZAB, GIS, IJS, KILJ, NLZOH, NUK, OILJ, SBCE, SBMB, UL, UM, UPUK
Spin-momentum locking is a direct consequence of bulk topological order and provides a basic concept to control a carrier's spin and charge flow for new exotic phenomena in condensed matter physics. ...However, up to date the research on spin-momentum locking solely focuses on its in-plane transport properties. Here, we report an emerging out-of-plane radiation feature of spin-momentum locking in a non-Hermitian topological photonic system and demonstrate a high performance topological vortex laser based on it. We find that the gain saturation effect lifts the degeneracy of the paired counterpropagating spin-momentum-locked edge modes enabling lasing from a single topological edge mode. The near-field spin and orbital angular momentum of the topological edge mode lasing has a one-to-one far-field radiation correspondence. The methodology of probing the near-field topology feature by far-field lasing emission can be used to study other exotic phenomena. The device can lead to applications in superresolution imaging, optical tweezers, free-space optical sensing, and communication.
Full text
Available for:
CMK, CTK, FMFMET, IJS, NUK, PNG, UL, UM
Previous studies suggest that gut microbiota is associated with neuropsychiatric disorders, such as Parkinson's disease, amyotrophic lateral sclerosis, and depression. However, whether the ...composition and diversity of gut microbiota is altered in patients with Alzheimer's disease (AD) remains largely unknown. In the present study, we collected fecal samples from 43 AD patients and 43 age- and gender-matched cognitively normal controls. 16S ribosomal RNA sequencing technique was used to analyze the microbiota composition in feces. The composition of gut microbiota was different between the two groups. Several bacteria taxa in AD patients were different from those in controls at taxonomic levels, such as Bacteroides, Actinobacteria, Ruminococcus, Lachnospiraceae, and Selenomonadales. Our findings suggest that gut microbiota is altered in AD patients and may be involved in the pathogenesis of AD.
Passive daytime radiative cooling (PDRC) involves spontaneously cooling a surface by reflecting sunlight and radiating heat to the cold outer space. Current PDRC designs are promising alternatives to ...electrical cooling but are either inefficient or have limited applicability. We present a simple, inexpensive, and scalable phase inversion-based method for fabricating hierarchically porous poly(vinylidene fluoride-co-hexafluoropropene) P(VdF-HFP)
coatings with excellent PDRC capability. High, substrate-independent hemispherical solar reflectances (0.96 ± 0.03) and long-wave infrared emittances (0.97 ± 0.02) allow for subambient temperature drops of ~6°C and cooling powers of ~96 watts per square meter (W m
) under solar intensities of 890 and 750 W m
, respectively. The performance equals or surpasses those of state-of-the-art PDRC designs, and the technique offers a paint-like simplicity.
Layered double hydroxide (LDH) is potentially excellent supercapacitor (SC) materials, but the low conductivity and easy agglomeration limit the further improvement of their electrochemical ...properties. Therefore, LDHs are requisite to grow on some conductive substrates to produce high-performance SC. In this paper, the conductive two-dimensional (2D) transition metal carbides, nitrides and carbonitrides (called MXene) were explored as the substrate to directly deposit NiFe-LDH nanosheets by a one-step hydrothermal method, then a three-dimensional (3D) porous NiFe-LDH/MXene electrode was obtained. The morphology and electrochemical performance of the composite electrodes were analyzed and investigated. The results show that the NiFe-LDH/MXene electrode has larger specific capacitance (720.2 F/g) than NiFe-LDH (465 F/g), and the capacitance of the composite electrode retained 86% after 1000 cycles (only 24% for NiFe-LDH), showing excellent cycle stability. The improved electrochemical performance of the composites is caused by the stable sheet-like structure of NiFe-LDH during charge-discharge time and the conductive network formed by the MXene, which can accelerates electron transport. In addition, the asymmetric SC based on NiFe-LDH/MXene positive electrode display a power density of 758.27 W/kg at an energy density of 42.4 Wh/Kg. These results indicate the NiFe-LDH/MXene composites can be applied as the novel candidate of high-performance SC electrodes.
•The LDH/MXene synergistic hybrid structure is developed with MXene nanosheets as the substrate.•Prominent electrical performance is attributed to MXene providing a fast charge transfer path for LDH.•The anchoring of LDH on MXene surface provides the material with enhanced cyclic stability.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP