Xinyu Zhu

SEP 10, 2026 · PREPRINT

Negative Self-Distillation: Learning to Reason by Avoiding Flaws

arXiv

This is an unrefereed preprint describing a novel machine learning method with empirical comparisons to baselines, but lacks peer review, independent validation, and clinical or real-world deployment evidence.

AUG 12, 2026 · JOURNAL ARTICLE

Predicting treatment-related cardiovascular risks in breast cancer patients: development and validation of an interpretable machine learning model

Frontiers in Oncology

A real result from a sound but limited retrospective study using machine learning on electronic health records; the model shows good discriminative performance (AUC 0.790) but lacks prospective validation and independent external testing.