Learning bipartite Bayesian networks under monotonicity restrictions

ABSTRACT

Learning parameters of a probabilistic model is a necessary step in machine learning tasks. We present a method to improve learning from small datasets by using monotonicity conditions. Monotonicity simplifies the learning and it is often required by users. We present an algorithm for Bayesian Networks parameter learning. The algorithm and monotonicity conditions are described, and it is shown that with the monotonicity conditions we can better fit underlying data. Our algorithm is tested on artificial and empiric datasets. We use different methods satisfying monotonicity conditions: the proposed gradient descent, isotonic regression EM, and non-linear optimization. We also provide results of unrestricted EM and gradient descent methods. Learned models are compared with respect to their ability to fit data in terms of log-likelihood and their fit of parameters of the generating model. Our proposed method outperforms other methods for small sets, and provides better or comparable results for larger sets.     

A Min–Max multiregulator system with stability analysis for aeroengine propulsion control

Stability analysis is an essential issue in Min–Max multiregulator control strategy for commercial aircraft engines. In this paper, a Min–Max selector scheme along with a stability analysis method is provided for aeroengine propulsion control. It is assumed that the main regulator is a dynamic compensator and the limit regulators are constant gains. The regulators are determined in such a way that the individual control loops are stable. However, due to the switching nature of Min–Max structure, the stability of single loops does not necessarily ensure the overall system stability. In order to analyze the stability of the presented Min–Maxapproach, the architecture of the control system is transformed into the canonical form of Lure’s system and the condition for absolute stability is specified using Multivariable Circle Criterion. The theoretical results can also be applied to investigate the stability of min-only or max-only schemes. Afterwards, using the provided methodology, the global asymptotic stability is proved for the control system of a high bypass two-spool turbofan engine and the performance of the designed Min–Max controller in tracking a desired fan speed and limit protection in fault-free and fault tolerant situations is compared with the well-known Min–Max/SMC approach.     

The relationship between oil prices and US economy revisited

The authors revisit the relationship between US economic growth and crude oil prices considering Industrial Production Index and West Texas Intermediate crude oil spot prices as respective proxies for a period spanning over January 1986 to June 2017. To capture the asymmetric and time-varying relationship, the authors employ maximum overlap discrete wavelet transform (MODWT)-based quantile regression (QR) analysis. Interestingly, MODWT-based QR analysis provides evidence of supply-driven link between crude oil prices and economic growth in the short run. However, in the medium to long run a demand-driven link is dominant. In addition, the QR results without MODWT also advocate a demand-driven link. Overall, the result of this study adds a new dimension to the literature on the relationship between crude oil prices and economic growth by focusing upon the time-frequency varying business cycle fluctuations.     

Improved Stable Indocyanine Green (ICG)‐Mediated Cancer Optotheranostics with Naturalized Hepatitis B Core Particles

In recent years, hepatitis B core protein virus‐like particle (HBc VLP) is an impressive biomaterial, which has attracted considerable attention due to favorable properties such as structural stability, high uptake efficiency, and biocompatibility in biomedical applications. Heretofore, only a few attempts have been made to apply it in physical, chemical, and biological therapy for cancer. In this study, a tumor‐targeting RGD‐HBc VLP is first fabricated through genetic engineering. For image‐guided cancer phototherapy, indocyanine green (ICG) is loaded into RGD‐HBc VLP via a disassembly/reassembly pathway and electrostatic attraction with high efficiency. The self‐assembled stable RGD‐HBc VLP significantly improves body retention (fourfold longer), aqueous stability, and target specificity of ICG. Remarkably, these positive reformations promote more accurate and sensitive imaging of U87MG tumor, as well as prolonged tumor destruction in comparison with free ICG. Moreover, the photothermal and photodynamic effect on tumors are quantitatively differentiated by multiple linear regression analysis. Overall, less‐potent medicinal ICG can be perfectly rescued by bioengineered HBc VLP to realize enhanced cancer optotheranostics.     

Development of surrogate models for the prediction of the flow around an aircraft propeller

In the present work, the derivation of two surrogate models (SMs) for modelling the flow around a propeller for small aircrafts is presented. Both methodologies use derived functions based on computations with the detailed propeller geometry. The computations were performed using k–ω shear stress transport for modelling turbulence. In the SMs, the modelling of the propeller was performed in a computational domain of disk-like geometry, where source terms were introduced in the momentum equations. In the first SM, the source terms were polynomial functions of swirl and thrust, mainly related to the propeller radius. In the second SM, regression analysis was used to correlate the source terms with the velocity distribution through the propeller. The proposed SMs achieved faster convergence, in relation to the detail model, by providing also results closer to the available operational data. The regression-based model was the most accurate and required less computational time for convergence.     

Artificial neural networks as alternative tool for minimizing error predictions in manufacturing ultradeformable nanoliposome formulations

This work was aimed at determining the feasibility of artificial neural networks (ANN) by implementing backpropagation algorithms with default settings to generate better predictive models than multiple linear regression (MLR) analysis. The study was hypothesized on timolol-loaded liposomes. As tutorial data for ANN, causal factors were used, which were fed into the computer program. The number of training cycles has been identified in order to optimize the performance of the ANN. The optimization was performed by minimizing the error between the predicted and real response values in the training step. The results showed that training was stopped at 10?000 training cycles with 80% of the pattern values, because at this point the ANN generalizes better. Minimum validation error was achieved at 12 hidden neurons in a single layer. MLR has great prediction ability, with errors between predicted and real values lower than 1% in some of the parameters evaluated. Thus, the performance of this model was compared to that of the MLR using a factorial design. Optimal formulations were identified by minimizing the distance among measured and theoretical parameters, by estimating the prediction errors. Results indicate that the ANN shows much better predictive ability than the MLR model. These findings demonstrate the increased efficiency of the combination of ANN and design of experiments, compared to the conventional MLR modeling techniques.     

Variability in rheology of cemented paste backfill with hydration age,binder and superplasticizer dosages

The use of (CPB) material to ameliorate geotechnical stability of underground mine is in nascent stage in India. Rheological properties of CPB change with travelling time as it is transported to underground mine stope through pipeline reticulation. In this paper, rheological properties of CPB based on mill tailings of a carbonate rich mineral processing waste are evaluated for different dosages of polycarboxylate (PC) based (SP). Each CPB sample having 78?wt% solids is mixed separately with 4%, 6% or 8% of binder dosages (ratio of the weight of dry binder to the weight of dry tailings) and, 0%, 0.5%, or 1.0% of SP dosages as weight of dry binder. The paper presents a methodology for determining yield stress, plastic viscosity and thixotropic behaviour of CPB mixture as a function of hydration age, binder and SP dosages. Results from the experimental campaigns indicate that SP content has significant influence on rheological behaviour of CPB and can be suitably exploited to enhance the flow characteristics of the carbonate rich process tailings. The study also develops multivariate linear regression models of yield stress, plastic viscosity and thixotropy of CPB depending on the hydration age, binder and SP dosages.     

Investigating volatile compounds’ contributions to the stale odour of green tea

This study aimed to characterise the volatiles that contribute to stale odour of green tea. Volatiles were extracted using headspace solid-phase microextraction (HS-SPME) and analysed using gas chromatography-mass spectrometry (GC-MS) and gas chromatography-olfactometry (GC-O). Results showed that a total of ninety-six volatiles were identified by GC-MS, in which forty-four volatiles were screened out based on Principal component analysis (PCA) and orthogonal projections to latent structures–discriminant analysis (OPLS-DA), and thirty-nine volatiles had a significant variation at the level of 0.05 by analysis of variance (anova ). From GC-O analysis, fifty-four aromatic volatiles with strong aroma intensity (aroma intensity above 2) were perceived. Further investigation revealed that fifteen volatiles, including 1-octen-3-ol, benzyl alcohol, benzaldehyde, safranal, β-cyclocitral, (E,E)-3,5-octadien-2-one, (Z,E)-3,5-octadien-2-one, dimethyl adipate, dihydroactinidiolide, β-ionone, α-ionone, geranyl acetone, phenylethyl alcohol, methyl decanoate and α-terpineol were responsible for stale odour, besides the former nine compounds were only found in stored tea.     

大型支撑辊的堆焊修复工艺及应用

大型支撑辊重量大、辊径大、辊面长、耐磨层厚,采用传统的堆焊工艺修复Cr3~Cr5大型支撑辊难度大,母体金属与堆焊层容易在堆焊时发生相变,导致开裂。另外,传统的轧辊堆焊修复工艺不合理,修复的轧辊力学性能较差,限制了轧辊的使用寿命,大量报废的轧辊尤其是大型轧辊长期堆积在轧钢厂内,增加了生产成本,造成了极大浪费。为此,开展了大型热轧Cr3~Cr5支撑辊及大型冷轧Cr3支撑辊焊材制备、堆焊工艺及修复技术研究。技术实施表明,修复后的支撑辊使用寿命达到新轧辊的寿命,每支堆焊修复支撑辊上机使用至报废尺寸,至少可循环堆焊修复3次,使支撑辊单项辊耗成本在每个循环周期内降低40%~60%,同时解决了大型支撑辊焊接性能不稳定的难题。