Closed certificates for fixed mirror-descent convergence (C73-15)

Closed-certificate trees and countable-ordinal rankings characterize convergence of fixed mirror-descent rules for the canonical regularizer and step-schedule class. The stated criterion is infinitary. Contents. Original proof PDF and LaTeX source, a complete research archive with pinned GitHub source snapshots and existing release assets, and a provenance manifest. The archive retains available verification records, source audits, reproducibility instructions and Lean source files. Original files are preserved byte-for-byte. Status. Author-submitted preprint and accompanying research materials. GitHub submission and Zenodo deposit do not constitute independent peer review or catalogue acceptance. This deposit records the existing results and their stated scope; archival checks do not certify mathematical correctness or replace a Lean build. AI use. The hand-proof manuscript and its supporting research materials were developed with GPT Codex assistance, including proof development and AI-assisted checks, as documented in the original repository. These checks are not represented as independent human peer review. Source versions. fixed-mirror-convergence-characterization — 000960f40cf5f7565032a9ba5a2986c93b5e0c5b Rights. CC BY 4.0 applies to the author-controlled materials that did not already carry a license. Existing file-level licenses, copyright notices and third-party attributions remain in force. Original manuscript and repository files have not been edited for this deposit.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-11
DOI
https://doi.org/10.5281/zenodo.23286419
Primary Topic
Stochastic Gradient Optimization Techniques
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

Closed certificates for fixed mirror-descent convergence (C73-15)

Ji Ho Bae
Zenodo (CERN European Organization for Nuclear Research)
Stochastic Gradient Optimization Techniques
preprint

Closed certificates for fixed mirror-descent convergence (C73-15)

Ji Ho Bae
preprint en

Abstract

Closed-certificate trees and countable-ordinal rankings characterize convergence of fixed mirror-descent rules for the canonical regularizer and step-schedule class. The stated criterion is infinitary. Contents. Original proof PDF and LaTeX source, a complete research archive with pinned GitHub source snapshots and existing release assets, and a provenance manifest. The archive retains available verification records, source audits, reproducibility instructions and Lean source files. Original files are preserved byte-for-byte. Status. Author-submitted preprint and accompanying research materials. GitHub submission and Zenodo deposit do not constitute independent peer review or catalogue acceptance. This deposit records the existing results and their stated scope; archival checks do not certify mathematical correctness or replace a Lean build. AI use. The hand-proof manuscript and its supporting research materials were developed with GPT Codex assistance, including proof development and AI-assisted checks, as documented in the original repository. These checks are not represented as independent human peer review. Source versions. fixed-mirror-convergence-characterization — 000960f40cf5f7565032a9ba5a2986c93b5e0c5b Rights. CC BY 4.0 applies to the author-controlled materials that did not already carry a license. Existing file-level licenses, copyright notices and third-party attributions remain in force. Original manuscript and repository files have not been edited for this deposit.

Zenodo (CERN European Organization for Nuclear Research)
Stochastic Gradient Optimization Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.