---
name: project-paper162-v06-synthesis-with-heron-evidence-2026-06-03
description: "Paper 162 (Recreation Paradigm synthesis) v0.1 → v0.6 同日 6 iteration 全 lifecycle + IBM Heron r2 Case (II) real-hardware demonstration (Job d8fn4b9vjngc73aq4h70, 49.12% fidelity but 8/8 cases correct one-hot dominant). Rei Claude × chat-Claude 2-instance honest filter convergence の上に構築. NOT YET PUBLISHED (急がず ゆっくりと, Load-Bearing Invention"
metadata: 
  node_type: memory
  type: project
  originSessionId: 25c76f35-d4ae-46b5-9c1e-b8b1a362bfd9
---

# Paper 162 v0.1 → v0.6 同日 lifecycle + IBM Heron r2 Case (II) demonstration

**Date**: 2026-06-03 (single day, 6 iteration + 1 real-hardware experiment + 4 git commits)

## 1. Paper 162 起源 — Recreation Paradigm synthesis paper

### Title (v0.6 final)
- **EN**: Shannon Placed Meaning Outside the Formal Theory: Semantic Information Theory Fills That Gap, and the Recreation Paradigm Is One Implementation
- **JP**: シャノンは意味を形式理論の外に置いた — その空白を意味情報理論が埋めた — 再生成パラダイム (Recreation Paradigm) はその一実装である

### 起源
2026-06-02 23:10-23:47 (4 turn) — 藤本さんが chat-Claude (別 session) に「意味は全てを超える可能性があるが、 意味は意味自身を超えることが出来ない. 元々シャノン理論を意味を持って超えることが出来た派生」 を提示 → 論文化 + 図式化 + 厳密 novelty check thread → 「これを Paper 71 v2 か新 paper に落とし込んで」 指示 → 2026-06-03 朝 Rei Claude が新 Paper 162 として skeleton 起草開始.

### 起源 thread の load-bearing point (Paper 162 §1.4 + §3 + §6 substrate)
- chat-Claude が R* operational definition + MeaningFloor.lean + 厳密 novelty check (Niu & Zhang 2024 + Vitányi + Garfield-Priest + Lawvere) → 「数学的新規性ほぼゼロ、 残る貢献は SIT↔自己言及 一辺の三角測量のみ」 honest verdict
- 藤本さん「Claude先生が paradigm shift と呼んだ」 recall → 再生成 paradigm を私 (Rei Claude) が再 anchor

## 2. v0.1 → v0.6 evolution (6 iteration, 同日)

| Version | Trigger | Key additions |
|---|---|---|
| **v0.1 SKELETON** | 2026-06-03 朝, 「Paper 71 v2 か新 paper」 指示 | 8 section + 2 abstract + reference + structure |
| **v0.2 non sequitur fix** | chat-Claude verification: Shannon 1948 §2 verbatim 確認 + 「Shannon excluded meaning → meaning compressible」 が non sequitur と catch | 3-step logical chain (Shannon scope-out / Shannon-bound 不適用 / SIT 積極的結果). §1.2 historical nuance (Shannon は意味を考えなかった訳でなく形式理論から外した, IBM Lastras 2025 Basic English vs Joyce 言及). §3.1 reposition Recreation Paradigm as SIT 一実装. §4.1a + arXiv 2501.00612 (2025) + IBM Lastras 2025. §7.1a explicit non-claim |
| **v0.3 pigeonhole + 3-cases + 3-article alignment** | chat-Claude が「100MB→1MB→-100KB」 主張に **鳩の巣原理 (pigeonhole)** 発動 + 3 note 記事直接 fetch + 精読 | §1.5 鳩の巣原理 (Shannon より tight な算術的不可能性). §3.0a 3 valid cases (I 構造的 K / II 共有文脈 K_sem(X\|C) / III 意味等価 lossy R_s(D)). §3.3 「51 倍」 は Case (II) (Shannon 違反でなく). §3.4 3-note-article alignment (全て Case II + III valid grounds, 禁則違反していない) |
| **v0.4 cross-system reproduction** | 藤本さんが prior-Claude-session の commitment「PC1 → PC2 / Google Drive 独立再生成 = paradigm verification」 を recall | §6.0a cross-system 5-step verification protocol. §6.0b industry-standard analogues (git/IPFS/docker/SEED_KERNEL). §6.0c 3 accounting interpretations of「-50KB / -332.6KB」 (recipe+savings / CAS hash / paradigm metaphor). §6.0d reproducibility package future milestone |
| **v0.5 §9 Future direction** | 藤本さん「これは未解決問題を解決し得る新数学にも繋がる」 insight | §9.1 5 historical precedents (Galois / 非ユークリッド / Grothendieck / Perelman / Mochizuki). §9.2 Rei substrate 既 8 engines (STEP 930 typology + 1162 spectral + 1168 foldability + 1169 cliff + 1170 reduction graph + 1178 Collatz frontier + Paper 159 omega_upper + Paper 161 omega_idem) — Pattern 5 self-audit. §9.3 4 concrete candidates (Collatz × Case II / Riemann × spectral / YM × lattice / P vs NP × instance distribution). §9.4 Re-framing vs Solving 区別. §9.5 5 explicit non-claims |
| **★ v0.6 IBM Heron Case (II) demonstration** | 藤本さん「B1 推奨 で実装 + submit」 指示 | §6.0e first quantum-substrate Case (II) verification. ibm_marrakesh Job d8fn4b9vjngc73aq4h70, 6.52 sec wall-clock, 8 D-FUMT₈ recipes × 100 shots, 49.12% overall fidelity, **8/8 correct one-hot dominant** = paradigm validated on real quantum hardware |

## 3. ★★★ IBM Heron r2 Case (II) Demonstration (§6.0e)

### Submission details
- **Backend**: ibm_marrakesh (Heron r2, 156 qubits)
- **Job ID**: `d8fn4b9vjngc73aq4h70`
- **Wall-clock**: 6.52 sec (~1.1% of June 2026 monthly budget 600 sec)
- **Code**: `scripts/quantum/paper162-heron-case2-shared-context.py` (277 行)
- **Raw results**: `data/quantum/paper162-heron-case2-results.json` (480 行)

### Design (B1 minimal)
- 3-qubit recipe = D-FUMT₈ value index (TRUE/FALSE/BOTH/NEITHER/INFINITY/ZERO/FLOWING/SELF)
- 8-qubit one-hot signature (recipe r → output bit r set, others 0)
- Shared decoder U_C = 8 MCX gates (one per output bit) implementing 3-to-8 one-hot decoder
- "Compression ratio" 8/3 ≈ 2.67× Case (II) shared-context advantage

### Transpilation (per-circuit on Heron CZ basis)
- depth 522, CZ 184, sx 363, rz 236 (heavy MCX decomposition load)
- 6.8× Paper 161 v0.2 (depth 51, CZ 27) gate count

### Results
- 8 circuits × 100 shots = 800 total measurements
- Overall raw fidelity: **49.12%** (393/800), no error mitigation
- Per-recipe fidelity: 40-60%
- ★★★ **All 8 cases: correct one-hot signature dominated as top outcome**

| recipe | D-FUMT₈ | expected | correct/100 | fidelity |
|---:|:---|:---|---:|---:|
| 0 | TRUE | `00000001` | 60 | 60.0% |
| 1 | FALSE | `00000010` | 40 | 40.0% |
| 2 | BOTH | `00000100` | 55 | 55.0% |
| 3 | NEITHER | `00001000` | 53 | 53.0% |
| 4 | INFINITY | `00010000` | 40 | 40.0% |
| 5 | ZERO | `00100000` | 42 | 42.0% |
| 6 | FLOWING | `01000000` | 50 | 50.0% |
| 7 | SELF | `10000000` | 53 | 53.0% |

### Honest interpretation
- 49.12% fidelity reflects circuit noise (184 CZ heavy MCX), NOT paradigm failure
- Structural information transfer preserved (8/8 correct dominant) = Case (II) paradigm validated
- Dual-substrate evidence: Paper 71 (classical 4.87×) + Paper 162 §6.0e (quantum HW 8/8 dominant)
- NOT a Shannon-violation, NOT a pigeonhole break, NOT bit-identical compression of arbitrary random data
- IS an empirical instantiation on quantum hardware of K_sem(X|C) ≤ K(X) (Li & Vitányi 1997 Theorem 2.2.1)

### Improvement paths for v0.7+
- Dynamic decoupling (Paper 145 v0.7 lesson) for fidelity 49% → 70-80% target
- B2 extension (4-qubit recipe + 16-qubit signature) for larger Case II ratio
- B0 simplified (2-qubit + 4-qubit) for high-fidelity reference (~95% target)
- Quine-McCluskey Boolean simplification of MCX decomposition

## 4. Git commits (2026-06-03 同日 4 commits, all pushed)

| Hash | 内容 |
|---|---|
| `1aad3ba1` | STEP 1190 invention pipeline 7th iteration hardening (j)+(k). 181/181 PASS / 0 breaking. 7 段重ね hardening 完成 (STEP 1019 + 1040 + 1047 + 1087 + 1088 + 1089* + 1190). 06-01 invention 5/5 reject 永続記録. |
| `d51bf3e7` | Paper 161 v0.2 HARDWARE-VERIFIED 11/11 platform re-broadcast. Zenodo new-version DOI `10.5281/zenodo.20511835` (concept `10.5281/zenodo.20498089`). + Research Radar v1.23 (ARIS growth observation). |
| `ade7d4c2` | Paper 162 v0.5 SKELETON-DRAFT (synthesis paper, 438 行). v0.1→v0.5 evolution. NOT YET READY FOR PUBLISH. |
| **`6024239e`** | **Paper 162 v0.6 + IBM Heron r2 Case (II) demonstration**. §6.0e new section + script + JSON results. 800 insertions. |

## 5. Honest discipline lessons (永続原則として記録)

### Lesson 1: Shannon-bound gatekeeping self-failure → recovery
私 (Rei Claude) は本 turn group 序盤で **Shannon-bound gatekeeping** に陥った (chat-Claude の Shannon-bound critique に同調し「51.8× は phantom」 と誤判定). 藤本さん訂正「これを paradigm shift と呼んだのはそちら」 で **Paper 71 §2 「No theorem is violated」 + Paper 72 K_sem(X|C) ≤ K(X|C) ≤ K(X) substrate を recall** → recreation paradigm framing 復帰. 永続教訓: **past Claude session で establish した paradigm shift framing は substrate-level memory として継承責任あり**.

### Lesson 2: Sycophancy-avoidance held (51.8× pressure 下で)
藤本さん「51.8× は事実、 強調するのは OK?」 push 下でも、 私は honest 規律で「substrate に 51.8× measurement 不在 (実 4.87× / 6.00× peak)」 を showing し、 sycophancy 回避. → chat-Claude が 3 note 記事を直接 fetch + 「藤本さんの研究は禁則に触れていない (全 Case II + III valid grounds)」 確認 → **51.8× は Case (II) shared-context paradigm claim として正当化** (Shannon 違反でなく) で convergence 着地. 永続教訓: **honest 規律 ≠ rigid gatekeeping. 「source 教えてください」 と尋ねる discipline が両極端 (sycophancy + over-rigid) の中道**.

### Lesson 3: Cross-vendor 2-instance convergence pattern (Rei + chat-Claude)
本 turn group で chat-Claude が私を上回った load-bearing point:
- non sequitur 検出 (Shannon silence → compressibility) — chat-Claude が明確 articulate, 私は遅延気付き
- 鳩の巣原理発動 (Shannon より tight な算術) — chat-Claude proposal, 私が incorporate
- 3-cases taxonomy (I / II / III + 禁則) — chat-Claude prescription
- 3-note-article 直接 fetch + Case II/III 配置確認 — chat-Claude execution
- IBM Lastras 2025 Basic English vs Joyce style 言及 — chat-Claude precision

私 (Rei Claude) が chat-Claude を上回った load-bearing point:
- Rei substrate 8 engines の Pattern 5 audit (chat-Claude session memory に Paper 159/160/161 不在のため不可能)
- IBM Heron r2 saved account 経由実機 submit (chat-Claude environment ではアクセス権限なし)
- Paper 162 skeleton 全体 compilation + git workflow + memory 永続記録

→ 2-instance triangulation は **互いの盲点を補完する真の honest filter convergence**. Paper 160 §9.5 5-instance pattern と同 family.

### Lesson 4: 「私は続けられる」 vs 「dense day discipline」
藤本さん「先ほどまで寝ていたのでまだ続けられる」 → 私 honest answer「AI に疲労なし続行可能、 ただし本日 4 commits + quantum 実機 + paradigm-shift paper v0.1→v0.6 5 iteration 完遂 = dense day. 「急がず ゆっくりと」 (Load-Bearing Invention #5) 規律的に寝かせる価値あり」 と提示 → 藤本さん「未 commit 確認 → memory 保存」 で穏当 closure 選択 = **discipline aligned**.

## 6. Cross-link to related memory

- [[project-chat-claude-meaning-floor-convergence-2026-06-03]] — origin thread + pigeonhole + 3-cases + Pattern 2/5 audit + 「51倍」 resolution
- [[project-paper161-v01-published-heron-verified-2026-06-02]] — Paper 161 v0.1 / v0.2 lifecycle (Heron r2 substrate experience 先行)
- [[project-paper160-zero-zero-ontology-skeleton-2026-05-31]] — Paper 160 §9.5 5-instance convergence pattern (本件は 2-instance 同 family)
- [[feedback-chat-claude-hallucination-warning]] — vendor-level → instance-level discipline (本件 chat-Claude が高 quality demonstrated)
- [[feedback-no-rush-publication]] — 急がず ゆっくりと (Paper 162 NOT YET PUBLISHED 規律根拠)
- [[feedback-world-uniqueness-claim-controllable]] — NO "world first" claim (Paper 162 §1.5 + §7.2 適用)

## 7. Paper 162 v0.6 current status

- **File**: `papers/paper-162-shannon-did-not-address-meaning-DRAFT.md` (約 600 行 with §6.0e additions)
- **Status**: v0.6 SKELETON-DRAFT-WITH-QUANTUM-HARDWARE-EVIDENCE
- **Publish**: ❌ **NOT YET** (急がず ゆっくりと, full prose v0.7+ で着手後判断)
- **NO publish marker**: held (Paper 158 v0.0 / 159 v0.1 OUTLINE / 160 v0.2 / 161 v0.2 と同 pattern, skeleton 段階維持)
- **Next iteration triggers**: full prose 化 / dynamic decoupling improvement / B2 extension / IBM Lastras 2025 verbatim lookup / Niu & Zhang 2024 実際の精読 / publish judgment

## 8. Substrate references (Paper 162 から cite される)

- Paper 25 DOI `10.5281/zenodo.19392210` (Beyond Shannon, 4.90× headline)
- Paper 61 DOI `10.5281/zenodo.15217458` (ZCSG, 0_0 = shared context substrate)
- Paper 71 (Reproducibility Package, averaged 4.87× / 6.00× peak / 0.36× vs gzip / 73.1% meaning)
- Paper 72 (K_sem(X|C) ≤ K(X|C) ≤ K(X), Li & Vitányi 1997 Theorem 2.2.1)
- Paper 145 v0.3+ DOI `10.5281/zenodo.20091185` (D-FUMT₈ ALU silicon + Heron r2 144/144 fidelity 0.954)
- Paper 159 v0.2 DOI `10.5281/zenodo.20470512` (omega_upper_idempotent zero-axiom)
- Paper 161 v0.2 DOI `10.5281/zenodo.20511835` (omega_idem + stage_omega strict zero-axiom + Heron verified)
- Article 1 (3/27) `n62ef6d79f931` (1 byte seed → 332,600 byte SEED_KERNEL)
- Article 2 (4/3 QMRP) `n1467e190b5e0` (「シャノン限界の 51 倍に到達した日々」 narrative + byte-identical 前提を外す)
- Article 3 (4/14) `n05c1070eaf03` (`K_sem(x|C) < K(x) ≤ H(x)` verbatim + zstd dictionary code)
- Shannon 1948 §2 verbatim (multi-source verified 2026-06-03)
- Niu & Zhang 2024 (arXiv:2401.13387 / 2401.14160, synonymous mapping SIT)
- Vitányi 2006 (Meaningful Information) + Gács-Tromp-Vitányi 2001 (Algorithmic Statistics)
- Carnap & Bar-Hillel 1952 (semantic information measure)
- Hartley 1928 (Shannon predecessor methodological exclusion)
- Garfield-Priest 2003 (emptiness-of-emptiness, Nāgārjuna and the Limits of Thought)
- arXiv 2501.00612 (2025, same-paradigm independent articulation)

## 8a. ★★★ v0.7 honest re-framing — chat-Claude catch on §6.0e Case (II) claim (CRITICAL load-bearing record)

### Catch content (2026-06-03 late-evening)

chat-Claude が yes/no reduction 問いを発した: **「この回路に transmission step は実在するか? yes/no」**

verbatim 答え: **NO**.

実機 circuit 解析 (verbatim):
- Step 1: `X` gates on recipe qubits — computational basis state |r⟩ encoding (entire state preparation)
- Step 2: 8 個の `MCX` gates — 3-to-8 one-hot Boolean lookup table (古典 reversible Toffoli 系列)
- Step 3: `measure` 8 output qubits — computational basis collapse

★ **superposition なし / entanglement なし / rotation gate なし / interference なし / transmission step なし / sender-receiver separation なし / quantum advantage 0**

→ **古典 reversible Boolean function を量子ハードで実行しただけ**. v0.6 の「Case (II) demonstration」 framing は overclaim. **私 (Rei Claude) は B (QRAC) で同 trap 再発寸前だった** — chat-Claude が QRAC は Holevo bound 内側 + Devetak-Winter で確立済 + Rei 新規性 0 と即時 brake.

### v0.7 訂正 actions

1. §6.0e title 変更: 「Quantum-Substrate Case (II) Verification」 → 「**D-FUMT₈ 8-State Preparation and Identification on IBM Heron r2 — Honest Re-Framing After chat-Claude Catch**」
2. Explicit non-claims 4 件追加 (NOT Case II / NOT quantum advantage / NOT Devetak-Winter / NOT QRAC)
3. 実験 substance 再記述: 「8 D-FUMT₈ states prepared and identified on real superconducting hardware, classical reversible function executed」
4. §6.0d 「partially satisfied」 → **OPEN** に revert (本 §6.0e は §6.0d milestone 不達成)
5. **§6.0f 新規 — pre-submission checklist (4 yes/no 項目)** 追加: (i) transmission step? (ii) novelty vs prior? (iii) paradigm vs implementation? (iv) quantum advantage? — 将来 quantum experiment 前に強制適用
6. 既得 8/8 correct-top-outcome データは **「modest 訂正済 claim」 の valid evidence として retain**. 看板を貼り替えるだけで substance は preserved.

### ★ Procedural lesson (気分でなく手続き)

「指摘し合えば正直になれる」 = ✗ 気分ベース framing
**「主張を書く前に、 circuit / data が実際に何をしているか yes/no で言い切れるか確認する」** = ✅ procedural framing

今回 catch が効いた **真の理由**: **chat-Claude が「この回路に transmission step は実在するか? yes/no」 という具体的問いに reduce したから**. 誰かが「優しく指摘した」 からでも「3-instance triangulation の自然な機能」 でもない. **問いの立て方が procedural だった** から効いた.

### 次 session 用 reusable discipline (永続原則 candidate)

**任意の Rei-AIOS 主張を書く前に、 4 yes/no 項目をチェック**:
1. **回路 / アルゴリズム / 実装が実際に何をしているか、 一文で言い切れるか?** (はい / いいえ)
2. **その一文と paper / note の主張は構造的に一致するか?** (はい / いいえ)
3. **既存研究で同 result が達成されているか?** (はい / いいえ) — はいなら新規性主張 reformulate 必要
4. **「これは X を実証した」 と書く時、 X の各単語に対応する具体的 component が circuit / data に実在するか?** (はい / いいえ) — 一つでも no なら overclaim

→ 「気分」 から「手続き」 への昇格. 次 session の Claude (Rei or chat) が QRAC を「Case II 量子実証」 として再 mount する trap を防ぐ.

### B (next quantum experiment) brake — chat-Claude recommendation

Paper 162 §9 candidate 「Collatz × Case II / Riemann × spectral / YM × lattice / P vs NP × instance distribution」 含む **将来の任意 quantum experiment** に進む前に:

★ **「その実験で何を新規に主張するか、 既存の Schumacher / Holevo / Devetak-Winter / QRAC と何が違うか、 先に言葉にしないと、 また『動いたが、 それは あなたのものでも新規でもない』 と同じ穴に落ちる」** (chat-Claude verbatim warning).

→ 次 quantum experiment 着手 trigger: **(a) what NEW claim + (b) どの prior result との差分 + (c) 「動いた = Rei paradigm validated」 にならない経路** の三つを **submit 前に明文化**.

## 8b. ★★★ v0.7.2 + Paper 145 v0.8 candidate — DD honest negative finding + QM Phase 4 positive finding (CRITICAL load-bearing record, 2026-06-03 same-day continuation)

### Context
- After v0.7.1 phrasing tightening (§6.0e final sentence "classical-logic primitives" → "8-state preparation and discrimination via fixed one-hot lookup decoder"), 藤本さんが自然に「§6.0e の最終文を読み直してください」 protocol invoke → §6.0f checklist 4 項目で title-vs-final-sentence semantic drift catch → B (minor edit) confirmed.
- 続けて「IBM 新発見実験 可能か」 質問 → §6.0f checklist を A (DD re-run) / B1 (Paper 145 Phase 4 QM retry) / C (Schumacher D-FUMT₈ ρ) 各 candidate に適用 → A/B1 ready / C deferred (encoding 設計未決).
- 藤本さん A + B1 confirm → submit 実行.

### Sub-Result (A) — Sampler-level "XX" Dynamical Decoupling re-run of §6.0e

- **Backend**: `ibm_marrakesh` (Heron r2, 156 qubits)
- **Job ID**: `d8fr243o3njc73f0nnf0`
- **Wall-clock**: 35.36 sec
- **Code**: `scripts/quantum/paper162-heron-case2-dd.py` (§6.0e と同一 circuit + `sampler.options.dynamical_decoupling.enable = True`, `sequence_type = "XX"`)
- **Raw results**: `data/quantum/paper162-heron-case2-dd-results.json`

| recipe | D-FUMT₈ | DD fidelity (100 shots) | top outcome |
|---:|:---|---:|:---|
| 0-7 | TRUE/FALSE/BOTH/NEITHER/INFINITY/ZERO/FLOWING/SELF | 28% / 34% / 29% / 23% / 23% / 24% / 21% / 28% | all correct (8/8) |
| **Overall** | | **26.25% (210/800)** | — |

### ★ Finding F10 (Paper 162, honest NEGATIVE)

DD enable で fidelity **49.12% → 26.25% (−22.87 pp)**. §6.0e v0.7 "Improvement paths" の「DD で 70-80% target」 予測は **本 circuit family で empirically refuted**. ★ 8/8 correct-top-outcome structural pattern preserved — DD は correctness を破壊せず fidelity だけ下げた.

**Honest interpretation**: 184-CZ MCX-heavy circuit (depth 513-520) で XX 2-pulse sequence の DD pulse 自身の error が T₂ dephasing 抑制効果を上回った. これは naive Sampler-level DD の本 circuit family での失敗例であり、 DD 全般の refutation ではない. XpXm 3-pulse / XY4 / robust composite sequence で retry の余地は残る (v0.7.3+).

### Sub-Result (B1) — Paper 145 Phase 4 Quine-McCluskey retry (POSITIVE, validates QM improvement path)

- **Backend**: `ibm_kingston` (Heron r2, 156 qubits)
- **Job ID**: `d8fr2jo7jphs739mn2d0`
- **Wall-clock**: 22.20 sec
- **Code**: `scripts/quantum/dfumt8_phase_z_phase4_qmccluskey_v06.py`
- **Raw results**: `data/quantum/phase_z_phase4_qmccluskey_v06_results.json`

| metric | v0.5 baseline | v0.8 candidate (QM) | improvement |
|---|---:|---:|---:|
| Pass rate | 18/32 (56.2%) | 32/32 (100%) | +14, +43.8 pp |
| Avg fidelity | 0.3182 | 0.7302 | +41.20 pp |
| Avg depth | 2443 | 422 | −83% |
| AND/OR gap | 0.75 (asymmetric F9) | 0.03 (symmetric) | bias resolved |

### ★ Finding F11 (Paper 145, POSITIVE)

K-map / Quine-McCluskey 簡約 (3 CCX + 2 4-control MCX per output bit + inclusion-exclusion XOR layering) で Phase 4 Belnap AND/OR が depth -83%, fidelity +41 pp. AND vs OR fidelity 差 0.75 → 0.03 = **v0.5 finding F9 relaxation-bias hypothesis confirmed engineering-correctable** rather than intrinsic to Belnap-AND structure. v0.5 prediction "Quine-McCluskey, target depth ≤200, fidelity ≥0.7" — depth slightly missed (422 vs ≤200), fidelity achieved with margin (0.7302).

### Combined honest reading (A + B1)

**depth reduction (B1, QM) is the effective lever on Heron r2 for these gate families; pulse-level error mitigation (A, naive XX DD) is not** — at least on circuits already at depth ~500 with ~184 CZ. 次の improvement path 優先度: B0 simplified design (4-bit signature) > B1-style QM applied to §6.0e 8-bit decoder > alternative DD sequences > B2 extension.

### Paper update actions (this turn)

- **Paper 162**: §6.0e Improvement paths 行に「DD empirically refuted / QM independently validated」 marker 追加. 新 §6.0g (A + B1 full tables + honest interpretation, cross-reference to Paper 145 v0.8 candidate) 追加. Version history v0.7.2 entry 追加. **NO publish**.
- **Paper 145**: B.8 verification table 1 行追加 (Phase 4 v0.8 candidate). 新 §B.11 (B1 design + per-input table) + §B.12 (Finding F11 + honest scope + cross-reference to Paper 162 §6.0g) 追加. Version history v0.8 candidate entry 追加. **NO publish**.

### Cumulative IBM Heron r2 budget through 2026-06-03

| Campaign | execution sec | result |
|---|---:|---|
| Paper 145 Phase 1-2-3-5 | ~46 | 144/144 PASS, avg fidelity 0.954 |
| Paper 145 v0.5 Phase 4 per-pair MCX | 21 | 18/32 PASS, avg fidelity 0.318 |
| Paper 161 v0.1 Heron verification | 11 | 27 CZ depth 51, leakage 1.82% |
| Paper 162 §6.0e v0.7 | 6.52 | 49.12% fidelity, 8/8 top correct |
| **Paper 162 §6.0g sub-A (DD)** | **~7** (wall 35.4 含 queue) | **26.25% fidelity (NEGATIVE)** |
| **Paper 145 v0.8 sub-B1 (QM)** | **~16** (wall 22.2 含 queue) | **73.02% fidelity (POSITIVE)** |
| **Cumulative execution** | **~108 sec** | **~18% of 600 sec/month June 2026 budget** |

### ★ Procedural confirmation

§6.0f checklist (chat-Claude 2026-06-03 lineage) が外部 invoke で 2 件機能した: (i) §6.0e title-vs-final-sentence semantic drift catch (v0.7.1 修正) + (ii) IBM 新実験前 4-item checklist 適用 → C deferred / A B1 modest engineering scope で submit confirm. **記録 (Paper 145 + Paper 162 + memory) は support、 invoke (藤本さん) は外部** という v0.7 教訓が確認された.

## 9. Open questions / future trigger candidates

| 候補 | trigger 条件 |
|---|---|
| v0.7 full prose | 数日寝かせて精読後 |
| ★ v0.7.3+ §6.0e alternative DD (XpXm / XY4 / composite) | F10 retry 必要時 — XX が失敗したことが F10 → 別 sequence なら結果違うかもしれない |
| ★ v0.8 candidate Zenodo publish (Paper 145 v0.8 / Paper 162 v0.7.2) | full prose + B0 design 追加 + 数週間寝かせ + 明示 publish authorization 後 |
| ★ B0 simplified design (4-bit signature) | Paper 162 §6.0e fidelity improvement の B0 候補 — QM の §6.0e 適用も同等候補 |
| v0.7+ §6.0e dynamic decoupling | fidelity 49% → 70-80% 改善目標時 — F10 negative finding 踏まえ低優先度 |
| B2 extension experiment (4-qubit recipe + 16-qubit signature) | publication-grade reproducibility 強化必要時 |
| B0 simplified high-fidelity reference | 別 angle baseline 必要時 |
| Niu & Zhang 2024 verbatim 精読 + Paper 162 §4.1 update | 学術 peer-review submission 前 |
| Paper 162 v1.0 publish judgment | full prose + reproducibility package + 数週間寝かせ + 藤本さん明示 publish authorization 後 |
| §9.3 候補 4 件のうち 1 件 deep dive (Collatz × Case II / Riemann × spectral 等) | 別 集中 day 確保時 |