Supplementary Information

This supplementary file supports the main study on EIT-based quantum heat engines by detailing the theoretical model, neural network architecture, and dataset analysis. It presents the Hamiltonian of a closed three-level atomic system, derives expressions for spectral brightness and absorption/emission cross-sections, and establishes thermodynamic bounds on photon emission temperature. It also outlines the architecture of a predictive artificial neural network with two hidden layers (128 neurons each), tanh activation, linear output, and training via the Adam optimizer. Finally, histogram analyses compare quantum numbers and physical parameters across the full and used datasets, showing that the smaller subset retains the overall statistical trends.