Timezone: »
Spotlight
Outlier-Robust Optimal Transport
Debarghya Mukherjee · Aritra Guha · Justin Solomon · Yuekai Sun · Mikhail Yurochkin
Optimal transport (OT) measures distances between distributions in a way that depends on the geometry of the sample space. In light of recent advances in computational OT, OT distances are widely used as loss functions in machine learning. Despite their prevalence and advantages, OT loss functions can be extremely sensitive to outliers. In fact, a single adversarially-picked outlier can increase the standard $W_2$-distance arbitrarily. To address this issue, we propose an outlier-robust formulation of OT. Our formulation is convex but challenging to scale at a first glance. Our main contribution is deriving an \emph{equivalent} formulation based on cost truncation that is easy to incorporate into modern algorithms for computational OT. We demonstrate the benefits of our formulation in mean estimation problems under the Huber contamination model in simulations and outlier detection tasks on real data.
Author Information
Debarghya Mukherjee (University of Michigan)
Aritra Guha (Duke University)
Justin Solomon (MIT)
Yuekai Sun (University of Michigan)
Mikhail Yurochkin (IBM Research AI)
Related Events (a corresponding poster, oral, or spotlight)
-
2021 Poster: Outlier-Robust Optimal Transport »
Tue. Jul 20th 04:00 -- 06:00 PM Room
More from the Same Authors
-
2023 : Slicing Mutual Information Generalization Bounds for Neural Networks »
Kimia Nadjahi · Kristjan Greenewald · Rickard Gabrielsson · Justin Solomon -
2023 Poster: Simple Disentanglement of Style and Content in Visual Representations »
Lilian Ngweta · Subha Maity · Alex Gittens · Yuekai Sun · Mikhail Yurochkin -
2022 Poster: Log-Euclidean Signatures for Intrinsic Distances Between Unaligned Datasets »
Tal Shnitzer · Mikhail Yurochkin · Kristjan Greenewald · Justin Solomon -
2022 Spotlight: Log-Euclidean Signatures for Intrinsic Distances Between Unaligned Datasets »
Tal Shnitzer · Mikhail Yurochkin · Kristjan Greenewald · Justin Solomon -
2021 : Model fusion via single-round FL »
Mikhail Yurochkin -
2021 Expo Talk Panel: Enterprise-Strength Federated Learning: New Algorithms, New Paradigms, and a Participant-Interactive Demonstration Session »
Laura Wynter · Nathalie Baracaldo · Chaitanya Kumar · Parijat Dube · Mikhail Yurochkin · Theodoros Salonidis · Shiqiang Wang -
2020 Poster: Model Fusion with Kullback--Leibler Divergence »
Sebastian Claici · Mikhail Yurochkin · Soumya Ghosh · Justin Solomon -
2020 Poster: Two Simple Ways to Learn Individual Fairness Metrics from Data »
Debarghya Mukherjee · Mikhail Yurochkin · Moulinath Banerjee · Yuekai Sun -
2019 Poster: Bayesian Nonparametric Federated Learning of Neural Networks »
Mikhail Yurochkin · Mayank Agarwal · Soumya Ghosh · Kristjan Greenewald · Nghia Hoang · Yasaman Khazaeni -
2019 Oral: Bayesian Nonparametric Federated Learning of Neural Networks »
Mikhail Yurochkin · Mayank Agarwal · Soumya Ghosh · Kristjan Greenewald · Nghia Hoang · Yasaman Khazaeni -
2019 Poster: Dirichlet Simplex Nest and Geometric Inference »
Mikhail Yurochkin · Aritra Guha · Yuekai Sun · XuanLong Nguyen -
2019 Oral: Dirichlet Simplex Nest and Geometric Inference »
Mikhail Yurochkin · Aritra Guha · Yuekai Sun · XuanLong Nguyen -
2018 Poster: Stochastic Wasserstein Barycenters »
Sebastian Claici · Edward Chien · Justin Solomon -
2018 Oral: Stochastic Wasserstein Barycenters »
Sebastian Claici · Edward Chien · Justin Solomon