Timezone: »

 
Poster
Tails of Lipschitz Triangular Flows
Priyank Jaini · Ivan Kobyzev · Yaoliang Yu · Marcus Brubaker

Thu Jul 16 06:00 AM -- 06:45 AM & Thu Jul 16 07:00 PM -- 07:45 PM (PDT) @ None #None

We investigate the ability of popular flow models to capture tail-properties of a target density by studying the increasing triangular maps used in these flow methods acting on a tractable source density. We show that the density quantile functions of the source and target density provide a precise characterization of the slope of transformation required to capture tails in a target density. We further show that any Lipschitz-continuous transport map acting on a source density will result in a density with similar tail properties as the source, highlighting the trade-off between the importance of choosing a complex source density and a sufficiently expressive transformation to capture desirable properties of a target density. Subsequently, we illustrate that flow models like Real-NVP, MAF, and Glow as implemented lack the ability to capture a distribution with non-Gaussian tails. We circumvent this problem by proposing tail-adaptive flows consisting of a source distribution that can be learned simultaneously with the triangular map to capture tail-properties of a target density. We perform several synthetic and real-world experiments to complement our theoretical findings.

Author Information

Priyank Jaini (University of Amsterdam)
Ivan Kobyzev (Borealis AI)
Yaoliang Yu (University of Waterloo)
Marcus Brubaker (Borealis AI)

More from the Same Authors