时间空间序列
Online spatio-temporal matching in stochastic and dynamic domains 1-s2.0-S0004370218302030-main
[1805.00734] Multiscale socio-ecological networks in the age of information
时间序列 时间空间
时间空间
《Detecting Regions of Maximal Divergence for Spatio-Temporal Anomaly Detection》B Barz, E Rodner, Y G Garcia, J Denzler [Friedrich Schiller University] (2018) O网页链接 view:O网页链接 GitHub:O网页链接
《Spatio-Temporal Neural Networks for Space-Time Series Forecasting and Relations Discovery》A Ziat, E Delasalles, L Denoyer, P Gallinari [Sorbonne Universits] (2018) O网页链接 view:O网页链接
今日焦点：部分观测环境下的空间记忆生成时序模型《Generative Temporal Models with Spatial Memory for Partially Observed Environments》
M Fraccaro, D J Rezende, Y Zwols, A Pritzel, S. M. A Eslami, F Viola [Technical University of Denmark & DeepMind] (2018) O网页链接 view:O网页链接
[1804.09401] Generative Temporal Models with Spatial Memory for Partially Observed Environments
Sparse Wide-Area Control of Power Systems using Data-driven Reinforcement Learning
Amirhassan Fallah Dizche, Aranya Chakrabortty, Alexandra Duel-Hallen
Comments: Submitted to IEEE CDC 2018. 8 pages, 5 figures
Subjects: Systems and Control (cs.SY)
arXiv:1804.09827 [pdf, other]
In this paper we present an online wide-area oscillation damping control (WAC) design for uncertain models of power systems using ideas from reinforcement learning. We consider that the exact small-signal model of the power system at the onset of a contingency is not known to the operator and use online measurements of the generator states and control inputs to recursively learn a state-feedback controller that minimizes a given quadratic energy cost. However, unlike conventional linear quadratic regulators (LQR), we intend our controller to be sparse, so its implementation reduces the communication costs. We, therefore, employ the gradient support pursuit (GraSP) optimization algorithm to impose sparsity constraints on the control gain matrix during learning. The sparse controller is thereafter implemented using distributed communication. We highlight various implementation, convergence, and numerical benefits versus challenges associated with the proposed approach using the IEEE 39-bus power system model with 1149 unknown parameters.
《Graph-Based Deep Modeling and Real Time Forecasting of Sparse Spatio-Temporal Data》B Wang, X Luo, F Zhang, B Yuan, A L. Bertozzi, P. J Brantingham [UCLA] (2018) O网页链接 view:O网页链接
《Bag of Recurrence Patterns Representation for Time-Series Classification》N Hatami, Y Gavet, J Debayle [Ecole Nationale Superieure des Mines de Saint-Etienne] (2018) O网页链接 view:O网页链接
Online spatio-temporal matching in stochastic and dynamic domains 1-s2.0-S0004370218302030-main
[1805.00731] Exploring Emoji Usage and Prediction Through a Temporal Variation Lens
E:\搜狗高速下载2 2017.10-2018\1805.00731 Exploring Emoji Usage and Prediction Through a Temporal Variation Lens.pdf
[1805.00731] Exploring Emoji Usage and Prediction Through a Temporal Variation Lens
E:\搜狗高速下载2 2017.10-2018\1805.00731 Exploring Emoji Usage and Prediction Through a Temporal Variation Lens.pdf
We introduce a dynamical spatio-temporal model formalized as a recurrent neural network for forecasting time series of spatial processes, i.e. series of observations sharing temporal and spatial dependencies. The model learns these dependencies through a structured latent dynamical component, while a decoder predicts the observations from the latent representations. We consider several variants of this model, corresponding to different prior hypothesis about the spatial relations between the series. The model is evaluated and compared to state-of-the-art baselines, on a variety of forecasting problems representative of different application areas: epidemiology, geo-spatial statistics and car-traffic prediction. Besides these evaluations, we also describe experiments showing the ability of this approach to extract relevant spatial relations.
Comments: accepted by: ICDM 2018 - IEEE International Conference on Data Mining series (ICDM)
Subjects: Learning (cs.LG); Machine Learning (stat.ML)
Journal reference: 2017 IEEE International Conference on Data Mining (ICDM), New Orleans, LA, 2017, pp. 705-714
DOI: 10.1109/ICDM.2017.80
Cite as: arXiv:1804.08562 [cs.LG]
(or arXiv:1804.08562v1 [cs.LG] for this version)
我们引入了一个动态的时空模型，形式化为一个递归神经网络，用于预测空间过程的时间序列，即共享时间和空间相关性的一系列观测值。 该模型通过结构化的潜在动态组件学习这些依赖关系，而解码器预测来自潜在表示的观察结果。 我们考虑这个模型的几个变体，对应于关于该系列之间的空间关系的不同的在先假设。 该模型经过评估并与最先进的基线进行比较，根据代表不同应用领域的各种预测问题：流行病学，地理空间统计和车流量预测。 除了这些评估之外，我们还描述了显示这种方法提取相关空间关系的能力的实验。
