Reproduce “FlowBatt: Flow Matching for Probabilistic Battery Degradation Prediction” 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
Hamidreza Eivazi, Eslam Sharaawy, Stefan Wittek, André Hebenbrock, Raphael Ginster, Steffen Blömeke
FlowBatt generates state-of-health (SOH) trajectories from early-cycle data, rather than predicting only an endpoint. It encodes capacity matrices and an SOH prefix, then uses flow matching to generate degradation paths; repeated samples represent uncertainty. Across five fixed-split benchmarks, it has the lowest mean trajectory RMSE, including 0.88 ± 0.03% on MIX-20, but intervals under-cover and unseen-condition generalization is untested.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.