Search papers, labs, and topics across Lattice.
This paper introduces the Swarm-Inspired Emergent Synchronizer (SIES), a novel graph-dynamical framework that learns local interaction rules to achieve desired global organization in collective behaviors. By utilizing signed source-target-conditioned attention within an explicit evolution model, SIES enables agents to synchronize and adapt across various network scales and dynamics without the need for retraining. Key results demonstrate that SIES not only outperforms baseline methods in synchronization control and robot locomotion but also excels in heterophilous node classification tasks, showcasing its versatility and generalizability.
SIES achieves rapid synchronization and adaptive coordination across diverse tasks and scales, outperforming traditional methods without retraining.
Collective behavior arises when locally interacting units produce coordinated global organization, from synchronization in dynamical systems to task-relevant information flow on graphs. The central challenge is not only to explain how collective behavior emerges, but to design local interaction rules that can produce desired global organization and generalize across graphs, dynamics and tasks.To address this challenge, we introduce the Swarm-Inspired Emergent Synchronizer (SIES), a graph-dynamical framework that learns generalizable local-interaction laws for controllable collective organization. Each node is an agent-like dynamical unit with a state and task cue, and signed source-target-conditioned attention acts as an adaptive coupling term inside an explicit evolution model. Therefore, SIES combines an explicit dynamical engine with local agent intelligence, similar to biological swarms. For synchronization control, SIES learns a generalizable coupling operator that produces prescribed synchronization patterns for CDSs across untrained network scales, target phase relations, and intrinsic node dynamics without retraining. The learned operator also reaches gait-related modes faster than three oscillator baselines and generalizes synchronization-driven locomotion to simulated multi-legged robots of different scales and a physical hexapod after leg disablement. For graph representation learning, SIES applies the same signed interaction principle to message passing and achieves the highest performance among the compared methods on heterophilous node-classification benchmarks. Together, these results position SIES as a generalizable and learnable graph-dynamical interaction framework with promise for synchronization control, adaptive robot coordination, and heterophilous graph representation learning.
NUS