AleynaCeyran/ParaQNN-Quantum-Decoherence: ParaQNN v1.0.0: Equation-Free Discovery of Open Quantum Systems

Official Implementation: ParaQNN Framework This repository hosts the official PyTorch implementation and supplementary materials for the manuscript: "Learning Quantum Decoherence via Paraconsistent Logic: An Equation-Free Neural Network Framework" Summary This work introduces ParaQNN, a novel neuro-symbolic architecture that integrates Paraconsistent Annotated Logic (PAL) with deep learning to characterize open quantum system dynamics. Unlike standard Physics-Informed Neural Networks (PINNs) that rely on predefined governing equations, ParaQNN treats environmental noise and decoherence as "contradictory evidence," enabling the autonomous discovery of physical laws from highly corrupted measurements without a priori Hamiltonian knowledge. Key Features Equation-Free Discovery: A robust framework for disentangling coherent unitary evolution from non-Markovian noise (e.g., Telegraph, $1/f$ Pink noise) using a dialetheist logical layer. Multi-Regime Validation: Pre-configured simulation suites for Rabi Oscillations, Lindblad Dissipation, and time-dependent Mixed Regimes. Rigorous Reproducibility: Deterministic data generation and training pipelines anchored to fixed random seeds (42–46) to ensure the exact replication of the reported figures. Comprehensive Benchmarking: Comparative analysis against state-of-the-art baselines, including PINNs (Incomplete/Known Physics), GANs, XGBoost, and Random Forest. Repository Contents src/: Core source code for the ParaQNN model and quantum master equation solvers. data/synthetic/: Canonical datasets used in the manuscript for benchmarking. configs/: Hyperparameter configuration files (.yaml) ensuring strict experimental consistency. notebooks/: Step-by-step reproduction tutorials. This release serves as the frozen codebase for peer review and publication.