Reweighted Data for Robust Probabilistic Models

By Yixin Wang, Alp Kucukelbir, David M. Blei. 2016

12 citations recorded by the source corpus.

Probabilistic models analyze data by relying on a set of assumptions. When a model performs poorly, we challenge its assumptions. This approach has led to myriad hand-crafted robust models; they offer protection against small deviations from their assumptions. We propose a simple way to systematically mitigate mismatch of a large class of probabilistic models. The idea is to raise the likelihood of each observation to a weight. Inferring these weights allows a model to identify observations that match its assumptions; down-weighting others enables robust inference and improved predictive accuracy. We study four different forms of model mismatch, ranging from missing latent groups to structure misspecification. A Poisson factorization analysis of the Movielens dataset shows the benefits of reweighting in a real data scenario.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Active Bias: Training More Accurate Neural Networks by Emphasizing High Variance Samples (2017, background): “On the other hand, to make models more robust to outliers, some approaches inject bias into the loss function in order to emphasize easier examples [37, 48, 27, 35].”
Bayesian weighted Mendelian randomization for causal inference based on summary statistics (2018, unknown): “Inspired by the reweighted probabilistic model (RPM) (Wang et al., 2016), in BWMR model, the probability density function p(γ̂, Γ̂|β,γ,w;σ2X ,σ2Y , τ 2) is given as: p(γ̂, Γ̂|β,γ,w;σ2X ,σ2Y , τ 2) = 1 Z N∏ j=1 p(γ̂j|γj;σ2Xj) N∏ j=1 p(Γ̂j|β, γj;σ2Yj , τ) wj , where Z is the normalizing factor.”