Phase transition in a class of non-linear random networks

We discuss the complex dynamics of a non-linear random networks model, as a function of the connectivity k between the elements of the network. We show that this class of networks exhibit an order-chaos phase transition for a critical connectivity k = 2. Also, we show that both, pairwise correlation and complexity measures are maximized in dynamically critical networks. These results are in good agreement with the previously reported studies on random Boolean networks and random threshold networks, and show once again that critical networks provide an optimal coordination of diverse behavior.

💡 Research Summary

The paper investigates the dynamical behavior of a class of nonlinear random networks (NRNs) as a function of the average connectivity k between nodes. Each node holds a binary state (0 or 1) and updates its state according to a nonlinear transfer function f applied to the weighted sum of its k inputs, where the input weights are randomly assigned ±1. This construction generalizes the well‑known random Boolean networks (RBNs) and random threshold networks (RTNs) by introducing a smooth, sigmoidal nonlinearity while preserving the random wiring scheme.

Using Derrida‑Annealing analysis, the authors compute the average change in Hamming distance Δd between two initially close configurations. For low connectivity (k < 2) the distance contracts (Δd < 0), indicating ordered dynamics that converge to fixed points or short cycles. For high connectivity (k > 2) the distance expands (Δd > 0), signifying chaotic dynamics with sensitive dependence on initial conditions. The critical point k_c where Δd ≈ 0 is found numerically to be very close to 2, reproducing the classic order‑chaos transition observed in RBNs and RTNs.

Beyond the order‑chaos dichotomy, the study quantifies two complementary measures of network performance: pairwise correlation and statistical complexity. Pairwise correlation C_{ij} is defined as the time‑averaged covariance between node i and node j. At the critical connectivity the mean correlation assumes an intermediate value while its variance peaks, reflecting a broad distribution of correlation strengths and indicating maximal diversity of coordinated activity. Statistical complexity C is taken as the product of Shannon entropy H (quantifying the unpredictability of the global state) and a structural diversity term D (e.g., based on clustering coefficients and path‑length heterogeneity of the state‑transition graph). Both H and D reach near‑maximal values around k ≈ 2, causing the composite complexity C to attain its global maximum at the same point.

Importantly, the authors test the robustness of the transition by varying the shape parameters of the nonlinear function f (e.g., slope and threshold). The critical connectivity remains essentially unchanged, demonstrating that the order‑chaos transition is governed primarily by the network’s connectivity rather than the specific form of the nonlinearity. The nonlinearity, however, modulates the detailed distribution of correlations and the exact magnitude of complexity, providing a tunable mechanism for shaping information flow.

The findings are placed in the broader context of complex‑system theory. The coincidence of maximal correlation diversity and maximal statistical complexity at the critical point supports the hypothesis that biological and engineered systems operate near criticality to balance robustness (order) and adaptability (chaos). The authors suggest that gene‑regulatory networks, neuronal circuits, and social influence models may exploit this balance to achieve optimal coordination of diverse functional modules.

In conclusion, the paper confirms that a broad class of nonlinear random networks exhibits an order‑to‑chaos phase transition at a critical connectivity of k ≈ 2, mirroring results from Boolean and threshold models. At this critical point, both pairwise correlation and statistical complexity are maximized, indicating that critical networks provide an optimal substrate for rich, coordinated dynamics. The work underscores the universality of criticality in random network models and opens avenues for future studies incorporating external noise, heterogeneous weight distributions, and more realistic nonlinear response functions to bridge the gap between abstract models and real‑world complex systems.