1. 我爱机器学习(2)Machine Learning Daily: 2017-02-06arXiv Paper Daily: Mon, 6 Feb 20172. 深度学习世界(2)【AI 工厂】Facebook 计算机视觉 Lumos平台，内容理解之上的图像技术干货 | 详解支持向量机（附学习资源） ...继续阅读 (11)

Neural and Evolutionary ComputingRobust Particle Swarm Optimizer based on ChemomimicryCasey Kneale,Karl S. BookshComments: To be revised for formatting and submitted as a Letters style paperSubjects:Neural and Evolutionary Computing (cs.NE)A particle swarm optimizer (PSO) loosely based on the phenomena ofcrystallization and a chaos factor which follows the complimentary errorfunction is described. The method features three phases: diffusion, directedmotion, and nucleation. During the diffusion phase random walk is the onlycontributor to particle motion. As the algorithm progresses the contribu ...继续阅读 (6)

1. 我爱机器学习(2)Machine Learning Daily: 2017-02-06arXiv Paper Daily: Mon, 6 Feb 20172. 深度学习世界(2)【AI 工厂】Facebook 计算机视觉 Lumos平台，内容理解之上的图像技术干货 | 详解支持向量机（附学习资源） ...继续阅读 (10)

1. 我爱机器学习(1)Machine Learning Daily: 2017-02-052. 大数据文摘(1)不幸的人各有不幸吗?文本分析流浪汉乞讨标语牌后发现的套路(附代码)3. CSDN大数据(1)必读丨2016年度最热门的开发工具 ...继续阅读 (7)

1. Hic Rhodus, hic salta(1)Owl of Minerva 的 Live －－ 解构亲密关系2. 36大数据(3)信息图：数据科学家/数据工程师/统计师的职位及工作比较数据脱敏(Data Masking)与大数据脱敏平台设计特普朗撼动世界背后的大数据风暴 ...继续阅读 (10)

Neural and Evolutionary ComputingDominance Move: A Measure of Comparing Solution Sets in Multiobjective OptimizationMiqing Li,Xin YaoComments: 23 pages, 10 figuresSubjects:Neural and Evolutionary Computing (cs.NE); Optimization and Control (math.OC)One of the most common approaches for multiobjective optimization is togenerate a solution set that well approximates the whole Pareto-optimalfrontier to facilitate the later decision-making process. However, how toevaluate and compare the quality of different solution sets remainschallenging. Existing measures typically require additional problem ...继续阅读 (10)

Neural and Evolutionary ComputingDominance Move: A Measure of Comparing Solution Sets in Multiobjective OptimizationMiqing Li,Xin YaoComments: 23 pages, 10 figuresSubjects:Neural and Evolutionary Computing (cs.NE); Optimization and Control (math.OC)One of the most common approaches for multiobjective optimization is togenerate a solution set that well approximates the whole Pareto-optimalfrontier to facilitate the later decision-making process. However, how toevaluate and compare the quality of different solution sets remainschallenging. Existing measures typically require additional problem ...继续阅读 (3)

Neural and Evolutionary ComputingRobust Order Scheduling in the Fashion Industry: A Multi-Objective Optimization ApproachWei Du,Yang Tang,Sunney Yung Sun Leung,Le Tong,Athanasios V. Vasilakos,Feng QianSubjects:Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)In the fashion industry, order scheduling focuses on the assignment ofproduction orders to appropriate production lines. In reality, before a neworder can be put into production, a series of activities known aspre-production events need to be completed. In addition, in real productionprocess, owing to various unce ...继续阅读 (18)

Neural and Evolutionary ComputingRobust Order Scheduling in the Fashion Industry: A Multi-Objective Optimization ApproachWei Du,Yang Tang,Sunney Yung Sun Leung,Le Tong,Athanasios V. Vasilakos,Feng QianSubjects:Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)In the fashion industry, order scheduling focuses on the assignment ofproduction orders to appropriate production lines. In reality, before a neworder can be put into production, a series of activities known aspre-production events need to be completed. In addition, in real productionprocess, owing to various unce ...继续阅读 (5)

Neural and Evolutionary ComputingSkip Connections as Effective Symmetry-BreakingA. Emin OrhanComments: 16 pages, 12 figures, 1 supplementary figureSubjects:Neural and Evolutionary Computing (cs.NE)Skip connections made the training of very deep neural networks possible andhave become an indispendable component in a variety of neural architectures. Asatisfactory explanation for their success remains elusive. Here, we present anexplanation for the benefits of skip connections in training very deep neuralnetworks. We argue that skip connections help break symmetries inherent in theloss landscapes ...继续阅读 (13)

Neural and Evolutionary ComputingSkip Connections as Effective Symmetry-BreakingA. Emin OrhanComments: 16 pages, 12 figures, 1 supplementary figureSubjects:Neural and Evolutionary Computing (cs.NE)Skip connections made the training of very deep neural networks possible andhave become an indispendable component in a variety of neural architectures. Asatisfactory explanation for their success remains elusive. Here, we present anexplanation for the benefits of skip connections in training very deep neuralnetworks. We argue that skip connections help break symmetries inherent in theloss landscapes ...继续阅读 (3)

Neural and Evolutionary ComputingPathNet: Evolution Channels Gradient Descent in Super Neural NetworksChrisantha Fernando,Dylan Banarse,Charles Blundell,Yori Zwols,David Ha,Andrei A. Rusu,Alexander Pritzel,Daan WierstraSubjects:Neural and Evolutionary Computing (cs.NE); Learning (cs.LG)For artificial general intelligence (AGI) it would be efficient if multipleusers trained the same giant neural network, permitting parameter reuse,without catastrophic forgetting. PathNet is a first step in this direction. Itis a neural network algorithm that uses agents embedded in the neural networkwhose task ...继续阅读 (8)

Neural and Evolutionary ComputingDesign of PI Controller for Automatic Generation Control of Multi Area Interconnected Power System using Bacterial Foraging OptimizationNaresh Kumari,Nitin Malik,A. N. Jha,Gaddam MalleshamComments: SCOPUS indexed Singapore journal, ISSN (Print): 2319-8613, ISSN (Online): 0975-4024, 8 pages, 11 Figures, 5 tablesJournal-ref: Int. J. Engineering & Tech.,8(6),2779-2786,2016Subjects:Systems and Control (cs.SY); Neural and Evolutionary Computing (cs.NE)The system comprises of three interconnected power system networks based onthermal, wind and hydro power genera ...继续阅读 (5)

Computer Vision and Pattern RecognitionPose Invariant Embedding for Deep Person Re-identificationLiang Zheng,Yujia Huang,Huchuan Lu,Yi YangSubjects:Computer Vision and Pattern Recognition (cs.CV)Pedestrian misalignment, which mainly arises from detector errors and posevariations, is a critical problem for a robust person re-identification (re-ID)system. With bad alignment, the background noise will significantly compromisethe feature learning and matching process. To address this problem, this paperintroduces the pose invariant embedding (PIE) as a pedestrian descriptor.First, in order to alig ...继续阅读 (22)

Computer Vision and Pattern RecognitionLearning an attention model in an artificial visual systemAlon Hazan,Yuval Harel,Ron MeirJournal-ref: IEEE International Conference on the Science of ElectricalEngineering (ICSEE) (2016)Subjects:Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)The Human visual perception of the world is of a large fixed image that ishighly detailed and sharp. However, receptor density in the retina is notuniform: a small central region called the fovea is very dense and exhibitshigh resolution, whereas a peripheral region around it has much ...继续阅读 (15)

Neural and Evolutionary ComputingDynamic Mortality Risk Predictions in Pediatric Critical Care Using Recurrent Neural NetworksM Aczon,D Ledbetter,L Ho,A Gunny,A Flynn,J Williams,R WetzelComments: 18 pages (5 of which in appendix), 9 figuresSubjects:Machine Learning (stat.ML); Neural and Evolutionary Computing (cs.NE); Dynamical Systems (math.DS); Quantitative Methods (q-bio.QM)Viewing the trajectory of a patient as a dynamical system, a recurrent neuralnetwork was developed to learn the course of patient encounters in thePediatric Intensive Care Unit (PICU) of a major tertiary care center. Da ...继续阅读 (12)

Neural and Evolutionary ComputingRegularizing Neural Networks by Penalizing Confident Output DistributionsGabriel Pereyra,George Tucker,Jan Chorowski,Łukasz Kaiser,Geoffrey HintonComments: Submitted to ICLR 2017Subjects:Neural and Evolutionary Computing (cs.NE); Learning (cs.LG)We systematically explore regularizing neural networks by penalizing lowentropy output distributions. We show that penalizing low entropy outputdistributions, which has been shown to improve exploration in reinforcementlearning, acts as a strong regularizer in supervised learning. Furthermore, weconnect a maximum ...继续阅读 (26)

Neural and Evolutionary ComputingFusion of Heterogeneous Data in Convolutional Networks for Urban Semantic Labeling (Invited Paper)Nicolas Audebert(Palaiseau, OBELIX),Bertrand Le Saux(Palaiseau),Sébastien Lefèvre(OBELIX)Comments: Joint Urban Remote Sensing Event (JURSE), Mar 2017, Dubai, United Arab Emirates. Joint Urban Remote Sensing Event 2017Subjects:Neural and Evolutionary Computing (cs.NE); Computer Vision and Pattern Recognition (cs.CV)In this work, we present a novel module to perform fusion of heterogeneousdata using fully convolutional networks for semantic labeling. We in ...继续阅读 (27)

1. 智能单元(1)Q&A 知乎Live：从AlphaGo看人工智能前沿技术2. 新智元(4)【实测】谷歌翻译特朗普就职演讲，正确率如何（附技术原理解析）2017年最值得关注的人工智能概念之“迁移学习”DeepMind黄士杰：Master 60局棋有两处人为失误迁移学习进展：单次学习能力达人类水平（附论文）3. 机器之心(4)深度 | 我们真的找对了方法？新研究为欧美脑计划敲响警钟招聘 | 加入百度语音，收获在人工智能领域的未来！一周论文 | Image Caption任务综述业界 | FDA批准首个云深度学习临床医疗应用平台Atrerys4. 大数据文摘(3)只有天才才能做数学吗？用Spark和DBSCAN对地理定位数据进行聚类这个boring的世界有什么好嗨的？就有！5. CSDN大数据(1)给老朋友们推荐几个大数据和机器学习方面的公众号6. 36大数据(2)移动金融app运营应该关注的9个指标深度评论 | 大数据金融风控大浪淘沙强弱渐分，长路漫漫投机者勿入 ...继续阅读 (12)

Neural and Evolutionary ComputingDeep Neural Networks – A Brief HistoryKrzysztof J. CiosComments: 14 pages, 14 figuresSubjects:Neural and Evolutionary Computing (cs.NE); Computer Vision and Pattern Recognition (cs.CV); Learning (cs.LG)Introduction to deep neural networks and their history.Optimized Spatial Partitioning via Minimal Swarm IntelligenceCasey Kneale,Dominic Poerio,Karl S. BookshComments: To be submitted to the Journal of TechnometricsSubjects:Methodology (stat.ME); Neural and Evolutionary Computing (cs.NE)Optimized spatial partitioning algorithms are the corner stone of manys ...继续阅读 (18)

1. 我爱机器学习(1)arXiv Paper Daily: Fri, 20 Jan 20172. 深度学习世界(2)KDnuggets 本月最受欢迎：5 个不容错过的机器学习项目深度 | 不要只看论文，缺乏工程实践才是深度学习研究的瓶颈 ...继续阅读 (18)

[英文版]Rules of Machine Learning: Best Practices for ML Engineering from Google.http://martin.zinkevich.org/rules_of_ml/rules_of_ml.pdf[中文版]谷歌机器学习应用的四十三条经验法则by 机器之心  ...继续阅读 (18)

1. 爱可可 – 爱生活(2)《Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks》L Mescheder, S Nowozin, A Geiger [MPI Tubingen & Microsoft Research Cambridge] (2017) http://t.cn/RM1K5hZ ​《今日学术视野(2017.01.19)》 http://t.cn/RM1K7tP ​ ...继续阅读 (15)

Neural and Evolutionary ComputingTemporal Overdrive Recurrent Neural NetworkFilippo Maria Bianchi,Michael Kampffmeyer,Enrico Maiorino,Robert JenssenSubjects:Neural and Evolutionary Computing (cs.NE); Dynamical Systems (math.DS)In this work we present a novel recurrent neural network architecturedesigned to model systems characterized by multiple characteristic timescalesin their dynamics. The proposed network is composed by several recurrent groupsof neurons that are trained to separately adapt to each timescale, in order toimprove the system identification process. We test our framework on ti ...继续阅读 (13)

Neural and Evolutionary ComputingThe Incredible Shrinking Neural Network: New Perspectives on Learning Representations Through The Lens of PruningNikolas Wolfe,Aditya Sharma,Lukas Drude,Bhiksha RajComments: 30 pages, 36 figures, submission to ICLR 2017Subjects:Neural and Evolutionary Computing (cs.NE); Learning (cs.LG)How much can pruning algorithms teach us about the fundamentals of learningrepresentations in neural networks? A lot, it turns out. Neural network modelcompression has become a topic of great interest in recent years, and manydifferent techniques have been proposed to address th ...继续阅读 (9)

Neural and Evolutionary ComputingAutomatic Knot Adjustment Using Dolphin Echolocation Algorithm for B-Spline Curve ApproximationHasan Ali Akyürek,Erkan Ülker,Barış KoçerComments: The Journal of MacroTrends in Technology and Innovation, Vol 4. Issue 1. 2016Subjects:Graphics (cs.GR); Neural and Evolutionary Computing (cs.NE)In this paper, a new approach to solve the cubic B-spline curve fittingproblem is presented based on a meta-heuristic algorithm called ” dolphinecholocation “. The method minimizes the proximity error value of the selectednodes that mea ...继续阅读 (18)

Neural and Evolutionary ComputingLIDE: Language Identification from Text DocumentsPriyank Mathur,Arkajyoti Misra,Emrah BudurSubjects:Computation and Language (cs.CL); Neural and Evolutionary Computing (cs.NE)The increase in the use of microblogging came along with the rapid growth onshort linguistic data. On the other hand deep learning is considered to be thenew frontier to extract meaningful information out of large amount of raw datain an automated manner. In this study, we engaged these two emerging fields tocome up with a robust language identifier on demand, namely LanguageIdentification ...继续阅读 (6)

Neural and Evolutionary ComputingSimplified Minimal Gated Unit Variations for Recurrent Neural NetworksJoel Heck,Fathi M. SalemComments: 5 pages, 3 Figures, 5 TablesSubjects:Neural and Evolutionary Computing (cs.NE); Machine Learning (stat.ML)Recurrent neural networks with various types of hidden units have been usedto solve a diverse range of problems involving sequence data. Two of the mostrecent proposals, gated recurrent units (GRU) and minimal gated units (MGU),have shown comparable promising results on example public datasets. In thispaper, we introduce three model variants of the minima ...继续阅读 (1)

Neural and Evolutionary ComputingOpenNMT: Open-Source Toolkit for Neural Machine TranslationGuillaume Klein,Yoon Kim,Yuntian Deng,Jean Senellart,Alexander M. RushComments: Report forthis http URLSubjects:Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)We describe an open-source toolkit for neural machine translation (NMT). Thetoolkit prioritizes efficiency, modularity, and extensibility with the goal ofsupporting NMT research into model architectures, feature representations, andsource modalities, while maintaining competitive perform ...继续阅读 (9)

Computer Vision and Pattern RecognitionSee the Glass Half Full: Reasoning about Liquid Containers, their Volume and ContentRoozbeh Mottaghi,Connor Schenck,Dieter Fox,Ali FarhadiSubjects:Computer Vision and Pattern Recognition (cs.CV)Humans have rich understanding of liquid containers and their contents; forexample, we can effortlessly pour water from a pitcher to a cup. Doing sorequires estimating the volume of the cup, approximating the amount of water inthe pitcher, and predicting the behavior of water when we tilt the pitcher.Very little attention in computer vision has been made to liquid ...继续阅读 (11)