Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
To enforce hard multi-agent constraints without requiring centralized runtime coordination, this work formulates multi-agent flow matching as a control-affine dynamical system with decoupled generative guidance (DeGG-Flow). The framework establishes theoretical feasibility conditions and finite-horizon convergence guarantees for both shared multi-agent and private neighbor-dependent requirements, accompanied by a Wasserstein bound on guidance-induced distributional shift. Evaluated on collaborative robotics and affordance-constrained scene synthesis, DeGG-Flow guarantees 100% satisfaction of coupled hard constraints while zero-shot generalizing to team sizes unseen during training.
Multi-agent generative models can now provably satisfy complex, coupled hard constraints without centralized runtime coordination—even zero-shot scaling to agent counts never seen during training.
Generative modeling is widely used for producing diverse objects from complex, multimodal distributions. However, its expressivity does not, in general, come with formal guarantees that the generated objects satisfy hard constraints or requirements. In multi-agent generation, this problem becomes more challenging because a hard requirement can depend on multiple agents, while each agent may need to determine its own guidance input without relying on the simultaneously computed guidance inputs of other agents. To this end, we introduce DeGG-Flow, a general framework for multi-agent flow matching with decoupled generative guidance. By representing the generative process as a control-affine dynamical system, we develop guidance conditions for two classes of coupled requirements: shared requirements whose satisfaction depends on multiple agents together, and private requirements associated with each individual agent dependent on its neighbors. For both classes, we establish feasibility conditions and finite-horizon convergence guarantees. We further derive a Wasserstein bound that characterizes the distributional deviation induced by the guidance. We demonstrate DeGG-Flow on multi-robot collaboration for crossing a spatial gap by reconfiguring the environment, and on multi-object scene generation with affordance requirements. Across both applications, DeGG-Flow directly generates objects that satisfy all corresponding hard requirements, including at team sizes unseen during training.