Filtering out mislabeled training instances using black-box optimization and quantum annealing

TL;DR

This paper introduces a novel method combining black-box optimization and quantum annealing to effectively identify and remove mislabeled instances from training datasets, improving model robustness and generalization.

Contribution

It presents a new scalable framework that integrates surrogate model-based black-box optimization with quantum annealing for noise removal in datasets.

Findings

Quantum annealing achieves faster optimization than simulated methods.

The method effectively prioritizes removal of high-risk mislabeled data.

Experimental results show improved dataset quality and model performance.

Abstract

This study proposes an approach for removing mislabeled instances from contaminated training datasets by combining surrogate model-based black-box optimization (BBO) with postprocessing and quantum annealing. Mislabeled training instances, a common issue in real-world datasets, often degrade model generalization, necessitating robust and efficient noise-removal strategies. The proposed method evaluates filtered training subsets based on validation loss, iteratively refines loss estimates through surrogate model-based BBO with postprocessing, and leverages quantum annealing to efficiently sample diverse training subsets with low validation error. Experiments on a noisy majority bit task demonstrate the method's ability to prioritize the removal of high-risk mislabeled instances. Integrating D-Wave's clique sampler running on a physical quantum annealer achieves faster optimization and higher-quality training subsets compared to OpenJij's simulated quantum annealing sampler or Neal's simulated annealing sampler, offering a scalable framework for enhancing dataset quality. This work highlights the effectiveness of the proposed method for supervised learning tasks, with future directions including its application to unsupervised learning, real-world datasets, and large-scale implementations.