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.
New method improves counterfactual distribution learning for high-dimensional outcomes.
We develop in detail the theory of c-projective geometry, a natural analogue of projective differential geometry adapted to complex manifolds. We realise it as a type of parabolic geometry and describe the associated Cartan or tractor connection. A Kaehler manifold gives rise to a c-projective structure and this is one…
Diffusion models adapt to data geometry through log-domain smoothing.
Adaptive sampling improves graph diffusion models by maintaining uniform information speed.
CWGD measures gradient diversity weighted by curvature, improving SGD convergence.
GABI learns geometry from diverse systems to improve Bayesian inference.
This research explains why SGD generalizes better than ADAM in deep learning.