BIOP01-63: MoFlow run-to-run 재현성 결과 draft_v2/ko 반입 - #19
Merged
kakyungkim merged 1 commit intoAug 11, 2026
Conversation
results/moflow_runtorun_null.md(PR #17, 판정 MOFLOW-REPRODUCIBLE)를 Results·Methods 본문에 반영. 근접-0 세 쌍(MV×MoFlow, MoFlow×CRAK, MoFlow×VAE)이 MoFlow 자체 불안정이 아니라 방법 간 진짜 근접-0 일치임을 재실행 천장(+0.9999, run1 vs run2 +1.0000)으로 뒷받침. "never established / uninterpretable / not separable" 캐비엇을 결과로 교체.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
무엇을 했는가
BIOP01-63(PR #17, results/moflow_runtorun_null.md, 판정 MOFLOW-REPRODUCIBLE)이 병합된 뒤 draft에는 아직 반영되지 않았던 부분을 채웠다. kkkim님이 PR #17 코멘트(2026-08-06)에서 "이제 draft 반영으로 넘어가셔도 됩니다"라고 확인한 작업이다.
어디를 고쳤는가
draft_v2.md/draft_v2_ko.md각 2곳(영/한 동시, 표현만 다르고 수치·근거는 동일):+0.9999], run1 vs run2 +1.0000)으로 교체. 근접-0 세 쌍(MV×MoFlow −0.0118, MoFlow×CRAK +0.0019, MoFlow×VAE +0.0019)의 절댓값 최댓값(0.0118)이 이 천장에 훨씬 못 미치므로, 세 쌍은 MoFlow 자신의 불안정이 아니라 방법 간 진짜 근접-0 일치로 읽을
수 있다는 결론으로 정리. 같은 천장이 MoFlow의 원본 대 셔플 값(+0.113)도 재실행 잡음 수준 밖에 있음을 뒷받침한다는 문장도 추가.
velocity_matrix_runtorun_null.md)·MoFlow(moflow_runtorun_null.md) 각각에 대해별도로 재실행 귀무가 확보됐다는 서술로 교체.
무엇을 안 건드렸는가
뒷받침하며, MoFlow 자체에 새 인과 주장을 추가하지 않았다 — claim-defensibility 게이트 대상 밖으로 좁게 잡았다.
results/velocity_matrix_audit.md등 봉인 감사 문서는 읽기만 하고 건드리지 않았다(moflow_runtorun_null.md 자체 규율).kkkim님 리뷰 부탁드립니다.