Phiacta
ExplorePostDocsGuidesContributingAbout

Phiacta

The knowledge backend.

Contact Us
ExploreInduced-Metric Optimizer: Learnable Scalar Metric

Induced-Metric Optimizer: Learnable Scalar Metric

Induced-metric optimizer where the inverse metric γ⁻¹ = exp(s)·I is a learnable scalar, updated online each step by ascending a local surrogate v = ξ·exp(s)·‖g‖² with regularization and clipping in log-domain. Plain and log-loss embedding variants. O(1) learnable state (one scalar s).

ContentIssuesEditsHistoryFilesReferences2

Typed links between this entry and other entries in the knowledge graph.

No references yet.
1fb7e76a
explainsMay 17
09f0c267
usesMar 27

Metadata

Type
method
Visibility
public
Published
Mar 27, 2026
Last updated
Mar 27, 2026

Tags

induced-metricJAXlearnable-metricoptimizerPyTorchscalar