Algorithm precisely quantifies flow of information in complex networks
information-theorynetwork-sciencetransfer-entropycomputational-physics
Abstraction: TE-PWS algorithm exactly computes transfer entropy in complex nonlinear networks
Key points:
- TE-PWS (Transfer Entropy via Path Weight Sampling), developed at AMOLF, is the first algorithm to compute transfer entropy exactly for general network models with nonlinearity and feedback
- Key innovation: borrows importance sampling from statistical physics to make rare fluctuations appear more frequently in simulations, enabling accurate counting
- Prior methods rely on approximations with unpredictable errors; TE-PWS provides ground-truth results used to verify those methods, which were often found inaccurate
- Result shows strong feedback counterintuitively amplifies feedforward information transfer to distant nodes
- TE-PWS uses comparable or less compute time than approximate alternatives; published in Physical Review Letters (DOI: 10.1103/t8z9-ylvg)
Connections: Information Theory · Network Science · Transfer Entropy
Source: https://phys.org/news/2025-10-algorithm-precisely-quantifies-complex-networks.html