148 Hits in 2.9 sec

Data-driven Weight Initialization with Sylvester Solvers [article]

Debasmit Das, Yash Bhalgat, Fatih Porikli
2021 arXiv   pre-print
In this work, we propose a data-driven scheme to initialize the parameters of a deep neural network.  ...  Our data-driven method achieves a boost in performance compared to random initialization methods, both before start of training and after training is over.  ...  To address these challenges, we propose an efficient gradient-free approach to data-driven weight initialization.  ... 
arXiv:2105.10335v1 fatcat:agyumysbmvdwpfgkp4ulex3nha

Software package for mosaic-Hankel structured low-rank approximation

Konstantin Usevich, Ivan Markovsky
2019 2019 IEEE 58th Conference on Decision and Control (CDC)  
ACKNOWLEDGEMENTS The authors would like to thank Mariya Ishteva, who contributed to the development of the factorization-based solver for structured low-rank approximation.  ...  The default initial approximation is computed from a kernel of a Sylvester subresultant matrix.  ...  Given measured data, prior knowledge about the data generating system, and approximation criterion, find optimal in the specified sense approximation of the measured data that is consistent with the given  ... 
doi:10.1109/cdc40024.2019.9028867 dblp:conf/cdc/UsevichM19 fatcat:fugw6hvtmnhqrksfbikqn5zhhy

Graph Kernels [article]

S.V.N. Vishwanathan, Karsten M. Borgwardt, Imre Risi Kondor, Nicol N. Schraudolph
2008 arXiv   pre-print
When the graphs are sparse, conjugate gradient solvers or fixed-point iterations bring our algorithm into the sub-cubic domain.  ...  Through extensions of linear algebra to Reproducing Kernel Hilbert Spaces (RKHS) and reduction to a Sylvester equation, we construct an algorithm that improves the time complexity of kernel computation  ...  grant No. 031U112F within the BFAM (Bioinformatics for the Functional Analysis of Mammalian Genomes) project, part of the German Genome Analysis Network (NGFN), and by NIH grant GM063208-05 "Tools and Data  ... 
arXiv:0807.0093v1 fatcat:s4zdxsd32fd4xmloezwbja6324

A New Algorithm for Convex Biclustering and Its Extension to the Compositional Data [article]

Binhuan Wang, Lanqiu Yao, Jiyuan Hu, Huilin Li
2021 arXiv   pre-print
On the other hand, the application of biclustering has not progressed in parallel with the algorithm techniques.  ...  The key step of our methods utilizes the Sylvester Equation to derive the ADMM algorithm, which is new to the clustering research.  ...  Thus, we choose row weights w l to sum to 1{ ? p and the column weights u k to sum to 1{ ? n. Next, we discuss two criteria for tuning parameter selection in a data-driven manner.  ... 
arXiv:2011.12182v2 fatcat:onylus2refexhmvanvlebuv5va

A New Framework for H_2-Optimal Model Reduction [article]

Alessandro Castagnotto, Boris Lohmann
2017 arXiv   pre-print
The paper ends with a brief discussion on how the idea behind the framework can be extended to approximate further system classes, thus showing that this truly is a general framework for interpolatory  ...  Note that this error estimation is very similar to the one in [18] , where data-driven approaches are used to generate a surrogate model from data acquired during IRKA.  ...  Step 2: H 2 optimization with respect to Σ µ As the Model Function Σ µ is a good approximation of Σ locally with respect to the initial interpolation data, we can run the H 2 optimization with respect  ... 
arXiv:1709.07270v1 fatcat:yfffi2ug3fasvdoqxo4fjlmt6a

Multi-dimensional signal approximation with sparse structured priors using split Bregman iterations [article]

Yoann Isaac, Quentin Barthélemy, Cédric Gouy-Pailler, Michèle Sebag, Jamal Atif
2016 arXiv   pre-print
Our well-founded approach yields same accuracy as the other algorithms at the state-of-the-art, with significant gains in terms of convergence speed.  ...  Firstly, a generic spatio-temporal regularization term is designed and used together with the standard ℓ_1 regularization term to enforce a sparse decomposition preserving the spatio-temporal structure  ...  "close" weights).  ... 
arXiv:1609.09525v1 fatcat:gtopgkahgjd37fpdarb7iallmy

Recent progress on variable projection methods for structured low-rank approximation

Ivan Markovsky
2014 Signal Processing  
Their application to system identification, approximate common divisor, and data-driven simulation problems is described in this paper and is illustrated by reproducible simulation examples.  ...  This paper gives an overview of recent progress in efficient local optimization algorithms for solving weighted mosaic-Hankel structured low-rank approximation problems.  ...  Acknowledgements Sections 6, 7.6, and A are based on joint work with K. Usevich. The research leading to these results has Theory, algorithms, and applications".  ... 
doi:10.1016/j.sigpro.2013.09.021 fatcat:ccqgdlyfjbdgjbm2gat6qgqi24

emgr—The Empirical Gramian Framework

Christian Himpe
2018 Algorithms  
Empirical Gramian are an extension to the system Gramians for parametric and nonlinear systems as well as a data-driven method of computation.  ...  The empirical Gramian framework - emgr - implements the empirical Gramians in a uniform and configurable manner, with applications such as Gramian-based (nonlinear) model reduction, decentralized control  ...  Yet, due to the data-driven nature of the empirical Gramians, even unstable systems or systems with inhomogeneous initial conditions are admissible.  ... 
doi:10.3390/a11070091 fatcat:f3kvq4e5bzhrresk64qqks7fbi

Scalable Incremental Nonconvex Optimization Approach for Phase Retrieval [article]

Ji Li, Jian-Feng Cai, Hongkai Zhao
2019 arXiv   pre-print
to find a good initial guess.  ...  Our proposed method overcomes the excessive computational cost of typical SDP solvers as well as the need of a good initialization for typical nonconvex methods.  ...  IncrePR outperforms all the other state-of-the-art solvers markedly with random initialization.  ... 
arXiv:1807.05499v3 fatcat:gv4vfoybjjftbbgqv5splsbfzi

Algorithms and software for U-Pb geochronology by LA-ICPMS

Noah M. McLean, James F. Bowring, George Gehrels
2016 Geochemistry Geophysics Geosystems  
For example, exciting advances are being driven by Laser-Ablation ICP Mass Spectrometry (LA-ICPMS), which allows for rapid determination of U-Th-Pb ages with 10s of micrometer-scale spatial resolution.  ...  The LA-ICPMS community is now faced with archiving these data with associated analytical results and, more importantly, ensuring that data meet the highest standards for precision and accuracy and that  ...  ., Sylvester and Ghaderi, 1997; Ko sler and Sylvester, 2003; Gehrels et al., 2008] .  ... 
doi:10.1002/2015gc006097 fatcat:4xvfgochhja45bpu22oquey72m

Sparsity-driven weighted ensemble classifier

Atilla Ozgur, Fatih Nar, Hamit Erdem
2018 International Journal of Computational Intelligence Systems  
In the proposed method, ensemble weights finding problem is modeled as a cost function with the following terms: (a) a data fidelity term aiming to decrease misclassification rate, (b) a sparsity term  ...  In this study, a novel sparsity-driven weighted ensemble classifier (SDWEC) that improves classification accuracy and minimizes the number of classifiers is proposed.  ...  Sylvester and Chawla [17] proposed differential evolution to find suitable weights for ensemble base classifiers.  ... 
doi:10.2991/ijcis.11.1.73 fatcat:q63y2wjadzedlm2wn7peoasry4

Drug-target interaction prediction using Multi Graph Regularized Nuclear Norm Minimization

Aanchal Mongia, Angshul Majumdar
2020 PLoS ONE  
original data.  ...  Previous works on Drug Target Interaction (DTI) prediction have shown that incorporating drug and target similarities helps in learning the data manifold better by preserving the local geometries of the  ...  More efficient solvers have also been proposed.  ... 
doi:10.1371/journal.pone.0226484 pmid:31945078 pmcid:PMC6964976 fatcat:nlqno73qs5daflgs554d2at2au

Fast Fusion of Multi-Band Images Based on Solving a Sylvester Equation

Qi Wei, Nicolas Dobigeon, Jean-Yves Tourneret
2015 IEEE Transactions on Image Processing  
By exploiting the properties of the circulant and downsampling matrices associated with the fusion problem, a closed-form solution for the corresponding Sylvester equation is obtained explicitly, getting  ...  Maximizing the likelihoods leads to solving a Sylvester equation.  ...  ML estimation is purely data-driven while Bayesian estimation relies on prior information, which can be regarded as a regularization (or a penalization) for the fusion problem.  ... 
doi:10.1109/tip.2015.2458572 pmid:26208345 fatcat:q6rvtzzbmbcyzn6le4u7o56owu

Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models [article]

Sam Bond-Taylor, Adam Leach, Yang Long, Chris G. Willcocks
2021 arXiv   pre-print
network f with parameters θ, that describes the transformation of a hidden state x(t) ∈ R D [27] , ∂x(t) ∂t = f (x(t), t, θ). (41) Starting from input noise x(t 0 ), an ODE solver can solve an initial  ...  With that said, JS-divergence is not perfect; if 0 mass is associated with a data sample in a maximum likelihood model, KLdivergence is driven to infinity, whereas this can happen with no consequence in  ... 
arXiv:2103.04922v2 fatcat:nivlg3whyjhadhwdl2tsh5yciy

Fast-convergent distributed coordinated precoding for TDD multicell MIMO systems

Rasmus Brandt, Mats Bengtsson
2015 2015 IEEE 6th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)  
The reweighted RCRM is based on semidefinite programming; for this we use Convex.jl [15] together with the Mosek solver (version, single-threaded) [16] .  ...  Weighted Sum Rate Maximization and OTA Overhead Given user priorities ↵ i k 0, and assuming per-BS sum power constraints, a weighted sum rate (WSR) maximization DL training Initial phase UL training  ... 
doi:10.1109/camsap.2015.7383835 dblp:conf/camsap/BrandtB15 fatcat:72ctebzsgngqbebasgtlmwi3ru
« Previous Showing results 1 — 15 out of 148 results