Our findings underscore the crucial role of interlayer link directionality in shaping the emergent behavior of multiplex systems, with potential implications for the design and control of such systems. Kudos to my co-authors: Mateo Bouchet, Alex Tejedor and Xiangrong Wang. (4/4)

In this regime, diffusion becomes dynamically arrested due to interlayer asymmetry, fragmenting the system into disconnected dynamical components & preventing it from converging to equilibrium. We also show that this transition can occur in synthetic and real-world systems 3/4

Previous works showed that directionality within layers can enhance diffusion (e.g., superdiffusion & prime regime). Now, we demonstrate analytically and numerically that asymmetric interlayer links can both reproduce these regimes and, remarkably, lead to a novel phenomenon. 2/4

Our findings underscore the crucial role of interlayer link directionality in shaping the emergent behavior of multiplex systems, with potential implications for the design and control of such systems. Kudos to my co-authors: Mateo Bouchet, Alex Tejedor and Xiangrong Wang. (4/4)

In this regime, diffusion becomes dynamically arrested due to interlayer asymmetry, fragmenting the system into disconnected dynamical components & preventing it from converging to equilibrium. We also show that this transition can occur in synthetic and real-world systems (3/4)

Previous works showed that directionality within layers can enhance diffusion (e.g., superdiffusion & prime regime). Now, we demonstrate analytically and numerically that asymmetric interlayer links can both reproduce these regimes and, remarkably, lead to a novel phenomenon. (2/4)

This work took more time & effort than expected, but it is finally out for good! 🎉

Big thanks to my collaborators Carlos Gracia Lázaro & @ymoreno.bsky.social — always a real pleasure to work together.

See the full article at: doi.org/10.1016/j.ch...

Our framework demonstrates a scenario for facilitating the representability of deep neural networks through learning the underlying mechanism, which aims to steer applications for predicting complex behavior that learnable physical rules can drive. Kudos to my co-authors! 4/4

We evaluate the performance of our framework in predicting typical critical behavior in spreading dynamics on various synthetic and real-world networks. Our results show that governing rules can be learned effectively and significantly improve prediction accuracy. 3/4

More specifically, the learned governing rules further refine and guide the representative learning of neural networks from a series of dynamic graphs. This combination enables knowledge-based prediction for the critical behavior of dynamical networked systems 2/4

Keynotes #CSS/France2025 : Hugues Bersini (UBL), @camcom.bsky.social (EHESS), @ymoreno.bsky.social (BIFI), @elsaarcaute.bsky.social (UCL)

Guests: Benjamin Fagard (CNRS), Agnieszka Rusinowska (CNRS), Guillaume Deffuant (INRAE)

Full program: conferences.css-fr.org?page_id=474