Steganography using a 3-player game

Image steganography aims to securely embed secret information into cover images. Until now, adaptive embedding algorithms such as S-UNIWARD or Mi-POD, were among the most secure and most often used methods for image steganography. With the arrival of deep learning and more specifically, Generative Adversarial Networks (GAN), new steganography techniques have appeared. Among them is the 3-player game approach, where three networks compete against each other. In this paper, we propose three different architectures based on the 3-player game. The first architecture is proposed as a rigorous alternative to two recent publications. The second takes into account stego noise power. Finally, our third architecture enriches the second one with a better interaction between embedding and extracting networks. Our method achieves better results compared to existing works Hayes and Danezis (2017), Zhu et al. (2018), and paves the way for future research on this topic.     

基于深度学习的传感器优化布置方法

针对传感器优化布置(optimal sensor placement,简称OSP)问题,提出了一种新的使用深度神经网络的解决方案,并以简化的桥梁形状的桁架结构中的振动测试传感器优化为例进行了验证。首先,选择一种传统的传感器优化布置方法,对自动化生成的大量不同的桁架结构分别进行传感器优化布置计算,将所得优化布置结果在进行数据预处理后构建出深度学习方法所需要的训练集与验证集;其次,使用Python语言和深度学习框架TensorFlow设计实现与本研究问题适配的深度神经网络模型并训练;然后,随机生成了新的桁架结构参数;最后,将深度神经网络输出的传感器布置结果和传统方法的计算结果进行了比较,验证了本研究方法的有效性以及在速度上、可移植性与可扩展性方面的性能优势。     

智能变电站设备管控大数据分析系统研究

结合数据挖掘、大数据技术在电力系统中的应用现状，分析和设计了一种智能变电站设备管控大数据分析系统。对变电站设备运行状态的大数据特征及基本框架进行了分析，并重点阐述了变电站设备运行状态大数据分析系统的数据集成与预处理、数据存储与处理、设备状态评估模型以及数据可视化展示。实验结果表明，所设计的大数据分析系统能够大幅度提升变电站设备数据分析智能化水平，为变电站管控稳定运行提供更加有效和实用性的技术支撑。     

人工监督神经网络在煤层裂缝发育区识别中的应用

针对在地震属性应用中难以精细识别煤层裂缝发育区边界的难题,开展了基于人工监督神经网络技术的煤层裂缝发育区应用研究。从地震资料中获取倾角导向体以提取高质量的地震属性,以多属性为指导进行人工拾取样点,并基于多层感知器进行神经网络机器训练学习,建立裂缝的最优属性集,拓展整体数据后获得裂缝概率体,从而识别划分出煤层裂缝发育区。该技术在山西阳泉新元矿区进行了应用,煤层裂缝发育区的识别结果明显优于属性直接识别,但勘探区内暂无钻井资料,预测效果还有待进一步验证。     

Collective reflection strategy for moderating conformity tendency and promoting reflective judgment performance

Reflective judgement is crucial for medical-related practitioners in dealing with controversial issues. However, the conformity phenomenon is likely to occur and interfere with reflective judgement learning during interactive activities. Effective strategies are required to moderate the conformity behaviour tendency (CBT) and improve reflective judgement performance (RJP). This study demonstrates two significant results: (a) Compared with the guided self-reflection learning strategy, the online collective reflection (OCR) learning strategy effectively weakened the learners' general CBT while dealing with professional controversial issues; and (b) a significantly negative correlation between the RJP achieved and the change of CBT in online environment was detected in the OCR group. The implications and potential applications in higher education were discussed. Further studies are needed to confirm the long-term effects and the extending application to other professional studies.     

基于卷积神经网络的葡萄叶片氮含量识别方法

为实现对葡萄叶片氮素含量快速、便捷的识别，在卷积神经网络VGG-16网络结构基础上，将数据增广后的图像按不同梯度划分进行模型训练，通过十折交叉验证法探究最佳的训练集与验证集分配比例，并构建4个不同深度的网络模型进行训练对比，采用全局平均池化代替全连接层约简网络参数量。训练结果表明，氮含量梯度设为0.70%、0.35%和0.175%时，室内简单背景识别准确率分别为85.9%、76.2%和71.1%；晴天室外复杂背景下识别准确率分别为44.6%、35.0%和30.4%。研究结果表明利用VGG-16建立的网络学习模型对葡萄叶片氮含量识别提供了一种新的便捷方法，对农业信息化和智能化技术应用具有一定促进作用。     

Adaptive neuro-fuzzy algorithm applied to predict and control multi-phase flow rates through wellhead chokes

A Takagi-Sugeno adaptive neuro-fuzzy inference system (TSFIS) model is developed and applied to a dataset of wellhead flow-test data for the Resalat oil field located offshore southern Iran, the objective is to assist in the prediction and control of multi-phase flow rates of oil and gas through the wellhead chokes. For this purpose, 182 test data points (Appendix 1) related to the Resalat field are evaluated. In order to predict production flow rate (Q_L) expressed as stock-tank barrels per day (STB/D), this dataset includes four selected input variables: upstream pressure (Pwh); wellhead choke sizes (D64); gas to liquid ratio (GLR); and, base solids and water including some water-soluble oil emulsion (BS&W). The test data points evaluated include a wide range of oil flow rate conditions and values for the four input variables recorded. The TSFIS algorithm applied involves five data processing steps: a) pre-processing, b) fuzzification, c) rules base and adaptive neuro-fuzzy inference engine, d) defuzzification, and e) post-processing of the fuzzy model. The developed TSFIS model for the Resalat oil field database predicted oil flow rate to a high degree of accuracy (root mean square error = 247 STB/D, correlation coefficient = 0.9987), which improves substantially on the commonly used empirical algorithms used for such predictions. TSFIS can potentially be applied in wellhead choke fuzzy controllers to stabilize flow in specific wells based on real-time input data records.     

R2D2: A scalable deep learning toolkit for medical imaging segmentation

Deep learning has gained a significant popularity in recent years thanks to its tremendous success across a wide range of relevant fields of applications, including medical image analysis domain in particular. Although convolutional neural networks (CNNs) based medical applications have been providing powerful solutions and revolutionizing medicine, efficiently training of CNNs models is a tedious and challenging task. It is a computationally intensive process taking long time and rare system resources, which represents a significant hindrance to scientific research progress. In order to address this challenge, we propose in this article, R2D2, a scalable intuitive deep learning toolkit for medical imaging semantic segmentation. To the best of our knowledge, the present work is the first that aims to tackle this issue by offering a novel distributed versions of two well-known and widely used CNN segmentation architectures [ie, fully convolutional network (FCN) and U-Net]. We introduce the design and the core building blocks of R2D2. We further present and analyze its experimental evaluation results on two different concrete medical imaging segmentation use cases. R2D2 achieves up to 17.5× and 10.4× speedup than single-node based training of U-Net and FCN, respectively, with a negligible, though still unexpected segmentation accuracy loss. R2D2 offers not only an empirical evidence and investigates in-depth the latest published works but also it facilitates and significantly reduces the effort required by researchers to quickly prototype and easily discover cutting-edge CNN configurations and architectures.     

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.