The Hierarchy of Stable Distributions and Operators to Trade Off Stability and Performance

05/27/2019
by   Adarsh Subbaswamy, et al.
0

Recent work addressing model reliability and generalization has resulted in a variety of methods that seek to proactively address differences between the training and unknown target environments. While most methods achieve this by finding distributions that will be invariant across environments, we will show they do not necessarily find the same distributions which has implications for performance. In this paper we unify existing work on prediction using stable distributions by relating environmental shifts to edges in the graph underlying a prediction problem, and characterize stable distributions as those which effectively remove these edges. We then quantify the effect of edge deletion on performance in the linear case and corroborate the findings in a simulated and real data experiment.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
05/27/2019

Should I Include this Edge in my Prediction? Analyzing the Stability-Performance Tradeoff

Recent work addressing model reliability and generalization has resulted...
research
02/20/2020

I-SPEC: An End-to-End Framework for Learning Transportable, Shift-Stable Models

Shifts in environment between development and deployment cause classical...
research
05/10/2023

Causal Information Splitting: Engineering Proxy Features for Robustness to Distribution Shifts

Statistical prediction models are often trained on data that is drawn fr...
research
01/23/2018

Stable gonality is computable

Stable gonality is a multigraph parameter that measures the complexity o...
research
08/30/2019

Composite likelihood methods for histogram-valued random variables

Symbolic data analysis has been proposed as a technique for summarising ...

Please sign up or login with your details

Forgot password? Click here to reset