NODE: R-FRANCHI.EU / STUDY NOTES

Study Notes

Antonio Franchi
← Return to the main page

Evolving notes developed while studying, conducting research, and preparing lectures. They are shared to make useful conceptual reconstructions, connections, and derivations available to others. They are not peer-reviewed publications or definitive teaching material. They may contain errors or incomplete arguments. Corrections and suggestions are welcome.

Intrinsic First-Order Optimization: Differentials, Gradients, Constraints, and Multipliers

A geometric reconstruction of first-order optimization centered on the differential as the intrinsic first-order object. The notes connect differentials, metric gradients, preconditioning, Newton directions, equality constraints, conormal spaces, Lagrange multipliers, tangent projection, and retractions.