litlas
ログインアカウント作成

Sinkhorn Distances: Lightspeed Computation of Optimal Transport

著者: Marco Cuturi - Conference on Neural Information Processing Systems, NeurIPS 2013 被引用: 5,613

Abstract. Optimal transportation distances are a fundamental family of pa-rameterized distances for histograms. Despite their appealing theoretical prop-erties, excellent performance in retrieval tasks and intuitive formulation, their computation involves the resolution of a linear program whose cost is prohibi-tive whenever the histograms ’ dimension exceeds a few hundreds. We propose in this work a new family of optimal transportation distances that look at transportation problems from a maximum-entropy perspective. We smooth the classical optimal transportation problem with an entropic regularization term, and show that the resulting optimum is also a distance which can be com-puted through Sinkhorn-Knopp’s matrix scaling algorithm at a speed that is several orders of magnitude faster than that of transportation solvers. We also report improved performance over classical optimal transportation distances on the MNIST benchmark problem. 1.

Information Theory andStatistical MechanicsInformation Theory and Statistical MechanicsDiagonal Equivalence toMatrices with Prescribe…Diagonal Equivalence to Matrices with Prescribed Row and Column SumsConcerning nonnegativematrices and doubly…Concerning nonnegative matrices and doubly stochastic matricesGeneralized IterativeScaling for Log-Linear…Generalized Iterative Scaling for Log-Linear ModelsHarmonic Analysis onSemigroupsHarmonic Analysis on SemigroupsOn the scaling ofmultidimensional…On the scaling of multidimensional matricesNetwork Flows: Theory,Algorithms, and…Network Flows: Theory, Algorithms, and Applications.Elements of InformationTheoryElements of Information TheoryThe Earth Mover'sDistance is the Mallows…The Earth Mover's Distance is the Mallows Distance: Some Insights from StatisticsAn Efficient EarthMover's Distance…An Efficient Earth Mover's Distance Algorithm for Robust Histogram ComparisonApproximate earthmover's distance in…Approximate earth mover's distance in linear timeFast and robust EarthMover's DistancesFast and robust Earth Mover's DistancesA sparse algorithm fordense optimal transportA sparse algorithm for dense optimal transportA Sparse MultiscaleAlgorithm for Dense…A Sparse Multiscale Algorithm for Dense Optimal TransportOn coupling particlefilter trajectoriesOn coupling particle filter trajectoriesMachine learning unifiesthe modeling of…Machine learning unifies the modeling of materials and moleculesOptimal-TransportAnalysis of Single-Cell…Optimal-Transport Analysis of Single-Cell Gene Expression Identifies Developmental Trajectories in ReprogrammingQuantum WassersteinGenerative Adversarial…Quantum Wasserstein Generative Adversarial NetworksLearning to SimulateComplex Physics with…Learning to Simulate Complex Physics with Graph NetworksA GeometricUnderstanding of Deep…A Geometric Understanding of Deep LearningThe Born Supremacy:Quantum Advantage and…The Born Supremacy: Quantum Advantage and Training of an Ising Born MachineAdversarially RobustRepresentations with…Adversarially Robust Representations with Smooth EncodersParticle GraphAutoencoders and…Particle Graph Autoencoders and Differentiable, Learned Energy Mover's DistanceEfficient OptimalTransport Algorithm by…Efficient Optimal Transport Algorithm by Accelerated Gradient DescentSinkhorn Distances:Lightspeed Computation…Sinkhorn Distances: Lightspeed Computation of Optimal Transport過去の参考文献中心の論文この論文を引用する論文古い新しい

ノードをクリックするとフォーカスを固定、空白をクリックすると本論文に戻ります。ホバーで一時的にプレビューできます。各ノードのページはタイトルから開けます。