Filters








378 Hits in 4.3 sec

Leveraging Spatial and Temporal Correlations in Sparsified Mean Estimation [article]

Divyansh Jhunjhunwala, Ankur Mallick, Advait Gadhikar, Swanand Kadhe, Gauri Joshi
2021 arXiv   pre-print
While most existing work on sparsified mean estimation is agnostic to the characteristics of the data vectors, in many practical applications such as federated learning, there may be spatial correlations  ...  We leverage these correlations by simply modifying the decoding method used by the server to estimate the mean.  ...  Acknowledgments This research was generously supported in part by the NSF CRII Award (CCF-1850029), the NSF CAREER Award (CCF-2045694), the Qualcomm Innovation fellowship and the CMU Dean's fellowship.  ... 
arXiv:2110.07751v1 fatcat:psf726pajbbgbfq6mp44lyi6pu

Sparsification and Filtering for Spatial-temporal GNN in Multivariate Time-series [article]

Yuanrong Wang, Tomaso Aste
2022 arXiv   pre-print
The proposed spatial-temporal graph neural network displays superior performances with respect to baseline approaches, with no graphical information, and with fully connected, disconnected graphs and unfiltered  ...  In contrast with existing sparsification methods adopted in graph neural network, our model explicitly leverage time-series filtering to overcome the low signal-to-noise ratio typical of complex systems  ...  The temporal and spatial component of the current architecture are designed to compute in parallel and combined in the end.  ... 
arXiv:2203.03991v1 fatcat:ufl4vpwsozdg5his3nyiacghwu

Recovering task fMRI signals from highly under-sampled data with low-rank and temporal subspace constraints

Mark Chiew, Nadine N. Graedel, Karla L. Miller
2018 NeuroImage  
We introduce a method for improving reconstruction fidelity using external constraints, like an experimental design matrix, to partially orient the estimated fMRI temporal subspace.  ...  The proposed method extends low-rank approximation methods for under-sampled fMRI data acquisition by leveraging knowledge of expected task-based variance in the data, enabling improvements in the speed  ...  Acknowledgments This research was funded by the Royal Academy of Engineering (MC) and the Wellcome Trust (KLM, 202788/Z/16/Z).  ... 
doi:10.1016/j.neuroimage.2018.02.062 pmid:29501875 pmcid:PMC5953310 fatcat:x4j3wk5phnbzdp6ea4kh67xypy

Forecaster: A Graph Transformer for Forecasting Spatial and Time-Dependent Data [article]

Yang Li, José M. F. Moura
2020 arXiv   pre-print
Spatial and time-dependent data is of interest in many applications.  ...  Based on the topology of the graph, we sparsify the Transformer to account for the strength of spatial dependency, long-range temporal dependency, data non-stationarity, and data heterogeneity.  ...  Then, it leverages graph CNN to capture spatial dependency and RNN to capture the temporal dependency within the data.  ... 
arXiv:1909.04019v5 fatcat:uq63oetxgva5zo2mtmx6uv5y2e

Multiscale algorithm for reconstructing videos from streaming compressive measurements

Jae Young Park, Michael B. Wakin
2013 Journal of Electronic Imaging (JEI)  
We leverage this tradeoff in proposing a multiscale reconstruction algorithm that alternates between video reconstruction and motion estimation as it produces finer resolution estimates of the video.  ...  Our analysis of the temporal complexity of videos reveals an interesting tradeoff between the spatial resolution of the camera, the speed of any moving objects, and the temporal bandwidth of the video.  ...  Like our work, the system in [26] can operate with frame-by-frame CS measurements and is intended to exploit spatial and temporal correlations in the video.  ... 
doi:10.1117/1.jei.22.2.021001 fatcat:7pwwxnsl3bf57i33kigygspbgi

Street-level Travel-time Estimation via Aggregated Uber Data [article]

Kelsey Maass, Arun V Sathanur, Arif Khan, Robert Rallo
2020 arXiv   pre-print
Estimating temporal patterns in travel times along road segments in urban settings is of central importance to traffic engineers and city planners.  ...  In this work, we propose a methodology to leverage coarse-grained and aggregated travel time data to estimate the street-level travel times of a given metropolitan area.  ...  In a set of articles [19, 25] , the authors leverage data from Bluetooth and GPS probe sensors for travel-time estimation and validation.  ... 
arXiv:2001.04533v1 fatcat:67mimrnksveynjnoseoa6roxr4

Kronecker Compressive Sensing

M. F. Duarte, R. G. Baraniuk
2012 IEEE Transactions on Image Processing  
While the CS literature has mostly focused on problems involving 1-D signals and 2-D images, many important applications involve multidimensional signals; the construction of sparsifying bases and measurement  ...  First, such matrices can act as sparsifying bases that jointly model the structure present in all of the signal dimensions.  ...  ACKNOWLEDGEMENTS Thanks to Ron DeVore for many valuable discussions and for pointing us to [30, 31] and to Justin Romberg for pointing us to the remark in [23] .  ... 
doi:10.1109/tip.2011.2165289 pmid:21859622 fatcat:akcrzq74h5efjco7irhmhlovau

Accelerating functional MRI using fixed-rank approximations and radial-cartesian sampling

Mark Chiew, Nadine N. Graedel, Jennifer A. McNab, Stephen M. Smith, Karla L. Miller
2016 Magnetic Resonance in Medicine  
Purpose: Recently, k-t FASTER (fMRI Accelerated in Spacetime by means of Truncation of Effective Rank) was introduced for rank-constrained acceleration of fMRI data acquisition.  ...  Both retrospective and prospectively under-sampled data were used to assess the fidelity of the enhancements to the k-t FASTER technique in resting and task-fMRI data.  ...  ACKNOWLEDGMENTS We thank Kamil Ugurbil, Essa Yacoub, Steen Moeller, and Eddie Auerbach for providing the multiband sequence, enabling the acquisition of the data in the retrospective simulations.  ... 
doi:10.1002/mrm.26079 pmid:26777798 pmcid:PMC4847647 fatcat:ddel7w7ofzbptpvzbcg2sl2fb4

Distributed Compressive Sensing Reconstruction via Common Support Discovery

Wei Chen, Miguel R. D. Rodrigues, Ian J. Wassell
2011 2011 IEEE International Conference on Communications (ICC)  
The proposed method exploits both the intra-sensor correlation and the inter-sensor correlation to reduce the number of samples required for recovering the original signals.  ...  An innovative feature of our method is using the Fréchet mean of the signals to discover the common support of their sparse representations in some basis.  ...  Although the actual physical environmental information such as temperature and humidity is compressible owning to the presence of temporal correlation and spatial correlation, traditional signal compression  ... 
doi:10.1109/icc.2011.5962798 dblp:conf/icc/0016RW11 fatcat:znbtyhq7pzcqfkba4bhemkmlxu

A Frechet Mean Approach for Compressive Sensing Date Acquisition and Reconstruction in Wireless Sensor Networks

Wei Chen, Miguel R. D. Rodrigues, Ian J. Wassell
2012 IEEE Transactions on Wireless Communications  
The novelty of the method relates to the use of the Fréchet mean of the signals as an estimate of their sparse representations in some basis.  ...  This crude estimate of the sparse representation is then utilized in an enhanced data recovering convex algorithm, i.e., the penalized ℓ1 minimization, and an enhanced data recovering greedy algorithm,  ...  However, it is in general difficult to conceive efficient compression schemes that account for both temporal and spatial correlation, duo to the distributed architecture of WSNs [4] .  ... 
doi:10.1109/twc.2012.081612.111908 fatcat:afrm6ur2v5e4vc4yodpdzqqucm

Sparsity Based Approaches for Distribution Grid State Estimation - A Comparative Study

Shweta Dahale, Hazhar Sufi Karimi, Kexing Lai, Balasubramaniam Natarajan
2020 IEEE Access  
For this purpose, a suite of sparsity-based approaches that exploit the correlation among states/measurements in spatial as well as temporal domains have been proposed recently.  ...  Specifically, the performance and complexity of spatial methods (1-D compressive sensing and matrix completion) and spatio-temporal methods (2-D compressive sensing and tensor completion) are compared  ...  Apart from exploiting spatial correlation, the existence of inherent spatio-temporal correlation in states and measurements can be leveraged using tensor completion algorithms.  ... 
doi:10.1109/access.2020.3035378 fatcat:lzoryb3vonb4nedewy2zzos5ki

PCI-MDR: Missing Data Recovery in Wireless Sensor Networks using Partial Canonical Identity Matrix

Neha Jain, Anubha Gupta, Vivek Ashok Bohara
2018 IEEE Wireless Communications Letters  
Researchers have exploited different characteristics of WSN data, such as low rank, spatial and temporal correlation for missing data recovery.  ...  For instance, correct rank estimation is required for exploiting the low-rank behaviour of WSNs, whereas correlation information among the nodes should be known for exploiting spatial correlation.  ...  In this method, a weighted combination of spatial and temporal estimation is considered.  ... 
doi:10.1109/lwc.2018.2882403 fatcat:suvu44e5xngmjnv3vnoe5fq37a

Optimized Node Selection for Compressive Sleeping Wireless Sensor Networks

Wei Chen, Ian J. Wassell
2016 IEEE Transactions on Vehicular Technology  
While conventional compressive sleeping WSNs only exploit the spatial correlation of SNs, the proposed approach further exploits the temporal correlation by selecting active nodes using the support of  ...  In this paper, we propose an active node selection framework for compressive sleeping wireless sensor networks (WSNs) in order to improve signal acquisition performance, network lifetime and the use of  ...  While compressive sleeping WSNs only leverage spatial correlations, temporal correlation is exploited in the proposed selection of active SNs.  ... 
doi:10.1109/tvt.2015.2400635 fatcat:46er3dquyfe7dg37ymzd2rfxbm

More with less

Liwen Xu, Xiaohong Hao, Nicholas D. Lane, Xin Liu, Thomas Moscibroda
2015 Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing - UbiComp '15  
Naïve applications of compressive sensing do not work well for common types of crowdsourcing data (e.g., user survey responses) because the necessary correlations that are exploited by a sparsifying base  ...  are hidden and non-trivial to identify.  ...  As discussed earlier, data correlation plays a key role in the data structure conversion. In some cases, there is obvious spatial and temporal correlation that CS can exploit.  ... 
doi:10.1145/2750858.2807523 dblp:conf/huc/XuHLLM15 fatcat:ghmi52pnrzdqjd53xwfmrtshmq

Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition [article]

Ziyu Liu, Hongwen Zhang, Zhenghao Chen, Zhiyong Wang, Wanli Ouyang
2020 arXiv   pre-print
In this work, we present (1) a simple method to disentangle multi-scale graph convolutions and (2) a unified spatial-temporal graph convolutional operator named G3D.  ...  The proposed G3D module leverages dense cross-spacetime edges as skip connections for direct information propagation across the spatial-temporal graph.  ...  correlations that are previously overlooked by modeling spatial and temporal dependencies in a factorized fashion.  ... 
arXiv:2003.14111v2 fatcat:3tz5oc3jwbalto2rjxjpybiqpe
« Previous Showing results 1 — 15 out of 378 results