This work introduces RUBRIC, a generator-agnostic filtering framework that formulates synthetic sample selection as a quality-over-quantity optimization problem and shows that RUBRIC improves F1-macro and recall while maintaining comparable ROC-AUC across several generators.
Abstract
Class imbalance poses a fundamental challenge in risk-sensitive applications such as fraud detection and medical diagnosis, where minority-class samples are scarce yet critical for accurate classification. Existing oversampling methods generate synthetic samples to rebalance class distributions; however, they often produce large numbers of low-quality candidates that distort decision boundaries or introduce artifacts, leading to overfitting and degraded generalization. In this work, we introduce RUBRIC, a generator-agnostic filtering framework that formulates synthetic sample selection as a quality-over-quantity optimization problem. RUBRIC ranks candidates using a realism-utility trade-off: realism is quantified by a learned discriminator that distinguishes real samples from synthetic samples, while utility captures proximity to the decision boundary through a concave margin-based scoring function. We show that, under mild regularity conditions, the proposed filtering strategy monotonically tightens the generalization bound for margin-based classifiers by jointly reducing distribution shift and suppressing near-negative tail contributions. Through extensive experiments on credit-card fraud detection and other imbalanced benchmarks, we demonstrate that RUBRIC improves F1-macro and recall while maintaining comparable ROC-AUC across several generators. We also provide explicit lambda-sensitivity analysis to show how users can recover AUPRC when ranking quality is prioritized.
Noise-Aware Adaptive-Difficulty Oversampling (NADOS) is proposed, which separates the assessment of seed trustworthiness from the allocation of synthesis effort and provides a practical strategy for noisy imbalanced classification.
Ednel Ashraff Misran, Syahid Anuar, Adam Mohd Khairuddin· International Journal of Adv...· 0 citations
A novel oversampling algorithm: the adaptive weighting–synthetic minority oversampling technique (AW-SMOTE), which combines the two perspectives of boundary tightness and local density and provides global sample enhancement support.
Bounded predictive influence and reliability-guided geometry as complementary mechanisms for imbalanced learning with uncertain labels are supported as complementary mechanisms for imbalanced learning with uncertain labels.
Mushir Akhtar, J. Akarsh, M. Tanveer et al.· arXiv.org· 0 citations
This study introduces a novel approach, named SDCGAN, which oversamples with a Conditional Generative Adversarial Network (CGAN) where the generator is fed with strengthened distribution information extracted from a curated set of minority samples.
Yu-Lin Zhang, Xiao-Zhe Wang, Kai-Wen Xue et al.· International Journal of Dat...· 0 citations
SafeVAE-GAN is proposed, a resampling framework that integrates a conditional β-Variational Autoencoder with Wasserstein GAN to generate high-quality minority samples and provides a robust and scalable solution for multi-class imbalanced learning across diverse domains.
Nhiem Ba Nguyen, Sinh Van Nguyen, B. Nguyễn· Vietnam Journal of Computer...· 0 citations
RoBell-RVFL is proposed, a robust and lightweight generalized bell random vector functional link network that redefines how randomized models handle class imbalance and noisy data and achieves adaptive control over sample contributions without sacrificing the closed-form learning efficiency of RVFL networks.
A. Rahaman, A. Quadir, M. Tanveer· 0 citations
Related blog posts
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026