Get Started
Home
Topics
Search
Library
Research questionHow can learned graph simulators preserve contact history for accurate long-horizon granular-dynamics rollouts?Granular contacts continually form, break, and rearrange, making interaction history difficult to retain in a changing graph. Losing that history can destabilize long rollouts and distort predicted deposit geometry and energy release.
AI
AI Memory
Machine Learning
Research Paper
Technology
Latest papersRecent research connected to this question, newest first.TRACE: Spatiotemporal Contact Memory Graph Network Simulator for Granular DynamicsThe source studies TRACE, which stores interaction history on contact edges and uses a physics-structured force decoder. Evidence is limited to 2D and 3D granular column-collapse benchmarks, with comparisons against node-based graph simulators and a material point method reference solver.research paper · Sep 2, 2026
Related questions
How can we simulate and backpropagate through long coupled dynamical systems without stepwise time recursion?How can learned models approximate trajectory similarity across multiple granularities without losing relative similarity rankings?How can representation-learning methods for dynamic heterogeneous graphs be compared across temporal granularities?How can process simulators preserve latent-state uncertainty when generating trajectories from partial event logs?