LiBaGS: Lightweight Boundary Gap Synthesis for Targeted Synthetic Data Selection

TL;DR

LiBaGS is a lightweight, generator-agnostic method for selecting synthetic data that improves training by targeting boundary regions and avoiding redundancy.

Contribution

It introduces a novel boundary-gap based selection method with a stopping rule, soft labels, and diversity to enhance model training with synthetic data.

Findings

LiBaGS outperforms classical oversampling and augmentation methods.

It effectively targets sparse, realistic decision boundary neighborhoods.

Experiments demonstrate improved accuracy with LiBaGS.

Abstract

Synthetic data is useful only when the added samples fill missing parts of the training distribution that matter for the downstream task. We introduce LiBaGS, a lightweight, generator-agnostic method for targeted synthetic training data selection. LiBaGS scores candidate synthetic samples by combining decision-boundary proximity, predictive uncertainty, real-data density, and support validity, so that selected samples are both informative and likely to remain on the real data manifold. We then use a boundary-gap allocation rule that targets sparse but realistic decision-boundary neighborhoods, rather than simply adding more data or selecting only the most uncertain candidates. LiBaGS also learns when enough synthetic samples have been added through a marginal-value stopping rule, assigns softer labels near ambiguous boundaries, and uses a diversity objective to avoid redundant near-duplicate selections. Experiments show that LiBaGS improves accuracy over classical oversampling, hard augmentation, uncertainty and density ablations, and targeted-generation selection criteria.