hamid tebyanian

High-quality randomness underpins modern cryptography, yet real quantum random-number generators (QRNGs) can betray subtle correlations introduced by hardware imperfections and environmental drift. We introduce a quantum-inspired deep-learning framework that detects such residual structure with sensitivities unattainable by clas- sical statistical batteries. The architecture intertwines a hierarchy of fractal-memory recurrent layers, a Kerr-oscillator bifurcation model that projects learned patterns into a quantum Hilbert space, and a topological loss based on persistent homology. Device-specific noise—beam-splitter imbalance, detector dark counts, homodyne electronic noise—is injected during training so that optimisation jointly minimises prediction error, topological divergence and loss of quantum fidelity while a reinforcement signal rewards rapid anomaly discovery. Public DV, CV and IBMQ datasets that satisfy all NIST and Diehard tests nevertheless yield bit- prediction accuracies between 0.72 and 0.83 and KL divergences of 3.9×10−2–5.2×10−2, demonstrating that “true” quantum randomness can harbour exploitable patterns unless carefully characterised and extracted.

This study delves into the validity of quantum mechanical operators in the context of quantum gravity, recognizing the potential need for their generalization. A primary objective is to investigate the repercussions of these generalizations on the inherent non-locality within quantum mechanics, as exemplified by Bell’s inequality. Additionally, the study scrutinizes the consequences of introducing a non-zero minimal length into the established framework of Bell’s inequality. The findings contribute significantly to our theoretical comprehension of the intricate interplay between quantum mechanics and gravity. Moreover, this research explores the impact of quantum gravity on Bell’s inequality and its practical applications within quantum technologies, notably in the realms of device-independent protocols, quantum key distribution, and quantum randomness generation.