MLLGCOJul 21, 2022

Correcting Model Bias with Sparse Implicit Processes

arXiv:2207.10673v21 citationsh-index: 8
AI Analysis

This addresses model selection issues for Bayesian ML practitioners, but it is incremental as it expands on existing SIP experiments.

The paper tackles the problem of model bias in Bayesian machine learning, where incorrect model assumptions can hinder prediction performance, and shows that Sparse Implicit Processes (SIP) can correct this bias by providing predictive distributions that better reflect the data, as demonstrated on synthetic datasets.

Model selection in machine learning (ML) is a crucial part of the Bayesian learning procedure. Model choice may impose strong biases on the resulting predictions, which can hinder the performance of methods such as Bayesian neural networks and neural samplers. On the other hand, newly proposed approaches for Bayesian ML exploit features of approximate inference in function space with implicit stochastic processes (a generalization of Gaussian processes). The approach of Sparse Implicit Processes (SIP) is particularly successful in this regard, since it is fully trainable and achieves flexible predictions. Here, we expand on the original experiments to show that SIP is capable of correcting model bias when the data generating mechanism differs strongly from the one implied by the model. We use synthetic datasets to show that SIP is capable of providing predictive distributions that reflect the data better than the exact predictions of the initial, but wrongly assumed model.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes