Fairness Hub Technical Briefs: Definition and Detection of Distribution Shift

TL;DR

This paper discusses the challenges of distribution shift in machine learning, emphasizing its impact on model performance across various applications and focusing on defining and detecting such shifts in educational prediction tasks.

Contribution

It provides a formal definition of distribution shift and explores methods for its detection specifically within educational prediction models.

Findings

Distribution shift can significantly degrade model performance.

Detection methods can identify shifts before they impact outcomes.

Focus on educational settings highlights domain-specific challenges.

Abstract

Distribution shift is a common situation in machine learning tasks, where the data used for training a model is different from the data the model is applied to in the real world. This issue arises across multiple technical settings: from standard prediction tasks, to time-series forecasting, and to more recent applications of large language models (LLMs). This mismatch can lead to performance reductions, and can be related to a multiplicity of factors: sampling issues and non-representative data, changes in the environment or policies, or the emergence of previously unseen scenarios. This brief focuses on the definition and detection of distribution shifts in educational settings. We focus on standard prediction problems, where the task is to learn a model that takes in a series of input (predictors) $X=(x_1,x_2,...,x_m)$ and produces an output $Y=f(X)$.