LGMay 7
A Closed-Form Upper Bound for Admissible Learning-Rate Steps in Belief-Space Dynamics
arXiv:2605.0674117.1
Predicted impact top 98% in LG · last 90 daysOriginality Incremental advance
AI Analysis
It offers a theoretical foundation for learning-rate selection in belief-space models, replacing tuning with a computable bound.
The paper derives a closed-form upper bound for admissible learning-rate steps in belief-space dynamics, showing that admissibility corresponds to contractivity in KL/Bregman geometry, and provides a formula rather than a heuristic.
Learning-rate steps are usually treated as hyperparameters. This paper isolates a local beliefspace calculation: when an update is modeled as a projected forward step on the probability simplex, admissibility means contractivity in the natural KL/Bregman geometry. Under this model, the upper bound of an admissible step is not a tuning slogan but a formula.