A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
A new gradient estimator for categorical distributions reduces bias and variance.
Sharp bounds on heat kernel derivatives on incomplete manifolds.
We introduce a family of pairwise stochastic gradient estimators for gradients of expectations, which are related to the log-derivative trick, but involve pairwise interactions between samples. The simplest example of our new estimator, dubbed the fundamental trick estimator, is shown to arise from either a) introducin…
Improved diffusion bridge sampling with rKL-LD loss.
New method tests tree models without causing computational pressure.