Multi-timescale time encoding for CNN prediction of Fenna-Matthews-Olson energy-transfer dynamics
This work addresses error accumulation in quantum dynamics simulations for researchers in quantum chemistry and materials science, offering a data-efficient method for predicting long-time dynamics in pigment-protein complexes, though it is incremental as it builds on existing CNN and encoding approaches.
The authors tackled the problem of error accumulation in machine learning simulations of open quantum dynamics by developing a non-recursive convolutional neural network that maps system parameters and a redundant time encoding to excitation-energy-transfer populations in the Fenna-Matthews-Olson complex, achieving an absolute relative error below 0.05 beyond 20 ps for stable long-time extrapolation.
Machine learning simulations of open quantum dynamics often rely on recursive predictors that accumulate error. We develop a non-recursive convolutional neural networks (CNNs) that maps system parameters and a redundant time encoding directly to excitation-energy-transfer populations in the Fenna-Matthews-Olson complex. The encoding-modified logistic plus $\tanh$ functions-normalizes time and resolves fast, transitional, and quasi-steady regimes, while physics-informed labels enforce population conservation and inter-site consistency. Trained only on $0\sim 7 ps$ reference trajectories generated with a Lindblad model in QuTiP, the network accurately predicts $0\sim100 ps$ dynamics across a range of reorganization energies, bath rates, and temperatures. Beyond $20 ps$, the absolute relative error remains below 0.05, demonstrating stable long-time extrapolation. By avoiding step-by-step recursion, the method suppresses error accumulation and generalizes across timescales. These results show that redundant time encoding enables data-efficient inference of long-time quantum dissipative dynamics in realistic pigment-protein complexes, and may aid the data-driven design of light-harvesting materials.