Tag
This paper introduces ODEWorld, a continuous-time latent world model using Physical-Time Flow (PT-Flow) that learns a latent velocity field parameterized by an ordinary differential equation, enabling arbitrary temporal resolution, backward prediction, and improved planning for video generation and robotic control.
This paper unifies twelve continuous-time generative models under mean-field game theory via a cost tuple, introduces MFGLab (a PyTorch library that auto-shares training loops and solvers), and proposes DI-Flow with differentiable entropy for better mode coverage.
This paper presents a theoretical framework for feedback-coupled memory systems operating in continuous time.
The paper identifies failure modes in Hamiltonian Generative Networks (HGN) that prevent temporal generalization to different step sizes in non-conservative environments, and proposes targeted fixes for stable dynamics prediction at variable temporal resolutions.
This paper proves two necessary conditions for optimal inference in a mesh of sovereign agents with irregular, non-stationary observations: an adaptive timescale and gap-dependence, which are satisfied only by liquid (continuous-time) networks.
This paper presents a theoretical framework for deep reinforcement learning in continuous environments, modeling it as a continuous-time stochastic process using stochastic control theory. The authors characterize an actor-critic algorithm's dynamics in the infinite width limit of two-layer networks, deriving an equation for infinitesimal changes in state distribution under a vanishingly small learning rate.
Proposes a continuity criterion for extending discrete-time causal prior-data fitted networks to continuous time using stochastic differential equations, introducing a taxonomy and fine-grid integration method that outperforms naive integration on irregular observation schedules.
This paper introduces Continuous-Time Distribution Matching (CDM), a method for few-step diffusion distillation that migrates from discrete to continuous optimization to improve visual fidelity and preserve fine details.