Reproduce “Bifurcations from Weight Space: JEPAWG on Parametric Dynamical Systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tobias Göbel, Miranda C.N. Cheng
JEPAWG tests whether trained forecaster weights encode known transitions in dynamical systems. The authors train networks across physical parameter values, learn a smooth parameter-to-weight map, and measure how rapidly its decoded weights change. In Lorenz dynamics, the response marks the Bykov T-point, where no fixed point changes stability; this indicates a trajectory-dependent diagnostic, not a universal bifurcation detector.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.