SEP 4, 2026 · PREPRINT
Impact of Data Loss in Postprocessing on Training and Inference of Quantum Neural Networks
arXiv
A case study identifying a data loss bug in production quantum machine learning software, with quantified impact on inference and training but no controlled experiments, peer review, or validation beyond two IBM backends.
Reported
Data loss rate (low end)85%
Data loss rate (high end)99.6%
Inference accuracy before data lo…0.94