Current study

Loading current research module…

Abstract

Optimization is not a switch that makes loss decrease. It is a discrete-time dynamical system whose geometry, memory, noise, and regularization determine the trajectory.

DOMAIN—

READING SESSION—

CURRICULUMDay — of an open horizon

RESEARCHED—

Loading the complete article…

Technical appendix A

The exact executable experiment

Not a rewritten snippet. This is the source displayed and executed by the local module.

About interactivity

GitHub Pages cannot execute the Python/Streamlit simulator. Run the local dashboard for live parameter controls; use this source for reproducible experiments.

Experiment sourceView complete Python source

Technical appendix B

Primary-source ledger

Paper metadata, official documentation, and lecture IDs checked against original sources on generation day.

Curriculum index

The mapped horizon is a queue,
not an ending.

— mapped lessons·— generated just in time·— verified sources used

Supporting foundations

The disciplines underneath the disciplines

Research history

Provenance is part of the lesson.

Corrections append to history; the learner’s past context is never silently rewritten.