Two new research papers introduce novel methods for learning the dynamics of complex physical systems using unlabeled data. The first paper, "Two-Parameter Flows for Learning Population Dynamics of Physical Systems," proposes a technique using two-parameter flows to learn time-dependent probability densities without trajectory information. The second paper, "Leveraging Gauge Freedom for Learning Non-Gradient Population Dynamics of Stochastic Systems," presents an algorithm called Non-Gradient Inference Flows (NGIF) that infers non-gradient population dynamics by parameterizing general vector fields, improving distributional accuracy over gradient-restricted methods. AI
RANK_REASON Two arXiv papers detailing novel research methodologies for learning physical system dynamics.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
