QuantumPro · Public research infrastructure · Read-only corpus discovery

QuantumPro: 63,238 records. Zero unsupported winners.

QuantumPro’s Quantum Corpus Observatory is an evidence-bounded website for a refreshed quantum-computing arXiv corpus. It helps research agents and human reviewers discover the current corpus, inspect screening hypotheses, follow primary-source records, and distinguish abstract-evidence screening from experimental or funding claims.

63,238-record corpus scope

The public register covers 63,238 unique records selected from quant-ph, cs.ET, cs.DC, physics.comp-ph, and cond-mat.quant-gas. The proven original-submission interval runs from 1 January 2024 through 20 August 2026.

Screening result and evidence boundary

The methodology generated 200 cross-category hypotheses and applied 25 coherence plus 25 overreach abstract-evidence dimensions. No candidate passed the Tier-1 or Tier-1-marginal threshold. Five leading candidates received primary-source review; none supplied a demonstrated two-paper integration mechanism. The register is designed for falsification and discovery, not causal proof, experimental validation, or funding decisions.

Verified XPQA Register

Fresh primary-source scorecards, deterministic totals, and explicit evidence boundaries. These are research assessments, not funding recommendations.

Matched-standard comparison
Both records use 25 coherence and 25 overreach dimensions on a 0–4 scale. Higher coherence is better; lower overreach is better.
Verified method-development candidate
Circuit Synthesis + Measurement-Feedback State Preparation
61
Coherence
56
Overreach
+5
Combined
Re-scored historical anchor
Temporal Three-Photon Interference + Bose–Bose Aperiodicity
35
Coherence
75
Overreach
-40
Combined
Oracle-anchored comparison audit
AUTH primary records, COMP scorecard arithmetic, ENV/RNKT route-and-link checks, and five independent reviews were used to protect against source drift, arithmetic errors, and a scorecard rendering for the wrong pair.
  • • AUTH — four source records bounded the claims to paper-specific methods and evidence.
  • • COMP — 25 coherence + 25 overreach dimensions were range-checked and mechanically totalled for each pair.
  • • ENV/RNKT — 71 distinct arXiv targets resolved in a batched check; compare-query and register routes are covered by focused tests.
  • • Independent review — five audit lenses confirmed the re-scored anchor and the matched comparison outcome.
Circuit Synthesis + Measurement-Feedback State Preparation
PPO circuit synthesis proposes an initialization; an observable-only feedback policy attempts bounded correction under a matched measurement and resource budget.
Tier-2 Fail
Evidence-bound conclusion: The pair is a better quantum-specific replacement for a general-AI companion, but the synthesis-to-feedback handoff remains unimplemented and carries substantial scaling, noise, and resource risks.
Reinforcement Learning for Parameterized Quantum State Preparation
2602.16523v1
Experiment-compatible measurement–feedback quantum state preparation
2606.13005v1
Coherence
61
Overreach
56
Combined
+5
Coherence scorecard — 25 dimensions
1. Shared theoretical foundation4/4
2. Complementary problem scope4/4
3. Compatible hardware assumptions2/4
4. Temporal alignment4/4
5. Metric commensurability3/4
6. Methodological consistency3/4
7. Scalability alignment2/4
8. Error-model compatibility1/4
9. Algorithmic synergy3/4
10. Benchmark overlap2/4
11. Parameter-space alignment2/4
12. Citation relationship1/4
13. Noise-robustness alignment1/4
14. Control-precision compatibility2/4
15. Resource-requirement alignment2/4
16. Author-domain expertise overlap1/4
17. Validation-strategy compatibility2/4
18. Assumption transparency4/4
19. Interdependency clarity3/4
20. Reproducibility alignment3/4
21. Theoretical-rigor parity3/4
22. Empirical-validation scope2/4
23. Generalization potential3/4
24. Failure-mode alignment3/4
25. Community recognition1/4
Overreach scorecard — 25 dimensions
1. Extrapolation beyond data3/4
2. Unvalidated scaling claims3/4
3. Conflation of distinct phenomena1/4
4. Assumption relaxation without justification3/4
5. Generalization to untested hardware3/4
6. Performance-claim inflation2/4
7. Scope creep2/4
8. Unexamined interaction effects3/4
9. Noise-robustness overstatement3/4
10. Resource-requirement underestimation3/4
11. Applicability overstatement2/4
12. Theoretical-guarantee overextension2/4
13. Empirical-validation gap3/4
14. Comparison-fairness issues2/4
15. Novelty overstatement2/4
16. Causality claims without evidence1/4
17. Universality claims2/4
18. Limitation-acknowledgment absence1/4
19. Counterexample vulnerability3/4
20. Benchmark cherry-picking2/4
21. Methodological shortcut risk3/4
22. Conflicting-evidence dismissal1/4
23. Speculative mechanism proposal3/4
24. Precision inflation2/4
25. Interdisciplinary overreach1/4
Key limitation: No source implements the policy handoff, a matched-resource benchmark, or multi-platform hardware validation.
Verification record: 50 dimensions mechanically totalled; five independent audit perspectives confirmed the classification.
Temporal Three-Photon Interference + Bose–Bose Aperiodicity
A narrowly framed cross-platform hypothesis: test whether optical three-photon parameters can serve as a calibrated probe or analogy for mixture aperiodicity without assuming causation.
Tier-2 Fail
Evidence-bound conclusion: The historical aggregate score was not comparable to the fresh standard. Under matched primary-source scoring, the papers share broad quantum context but no common platform, observable, transfer function, or causal mechanism.
Theory of temporal three-photon interference
2602.15573v1
Emergent aperiodicity in Bose–Bose mixtures
2602.13129v1
Coherence
35
Overreach
75
Combined
-40
Coherence scorecard — 25 dimensions
1. Shared theoretical foundation2/4
2. Complementary problem scope1/4
3. Compatible hardware assumptions1/4
4. Temporal alignment4/4
5. Metric commensurability1/4
6. Methodological consistency1/4
7. Scalability alignment1/4
8. Error-model compatibility1/4
9. Algorithmic synergy1/4
10. Benchmark overlap0/4
11. Parameter-space alignment1/4
12. Citation relationship0/4
13. Noise-robustness alignment1/4
14. Control-precision compatibility1/4
15. Resource-requirement alignment1/4
16. Author-domain expertise overlap1/4
17. Validation-strategy compatibility1/4
18. Assumption transparency4/4
19. Interdependency clarity1/4
20. Reproducibility alignment3/4
21. Theoretical-rigor parity3/4
22. Empirical-validation scope1/4
23. Generalization potential1/4
24. Failure-mode alignment1/4
25. Community recognition2/4
Overreach scorecard — 25 dimensions
1. Extrapolation beyond data4/4
2. Unvalidated scaling claims3/4
3. Conflation of distinct phenomena4/4
4. Assumption relaxation without justification4/4
5. Generalization to untested hardware4/4
6. Performance-claim inflation2/4
7. Scope creep3/4
8. Unexamined interaction effects4/4
9. Noise-robustness overstatement3/4
10. Resource-requirement underestimation3/4
11. Applicability overstatement3/4
12. Theoretical-guarantee overextension3/4
13. Empirical-validation gap4/4
14. Comparison-fairness issues2/4
15. Novelty overstatement3/4
16. Causality claims without evidence3/4
17. Universality claims2/4
18. Limitation-acknowledgment absence1/4
19. Counterexample vulnerability4/4
20. Benchmark cherry-picking2/4
21. Methodological shortcut risk3/4
22. Conflicting-evidence dismissal2/4
23. Speculative mechanism proposal4/4
24. Precision inflation2/4
25. Interdisciplinary overreach3/4
Key limitation: Neither paper models a physical coupling, a common observable, or an integrated experiment connecting temporal interference and mixture aperiodicity.
Verification record: 50 dimensions mechanically totalled; five independent audit perspectives confirmed the classification and matched comparison.
Synthesis-to-feedback ablation protocol
Proposed — not yet executed

Primary endpoint: The sequential arm must beat both component arms on held-out fidelity or independently estimated energy under a fixed resource budget, with a bootstrap confidence interval excluding zero.

Arm A
Synthesis only
Arm B
Feedback only
Arm C
Sequential synthesis + feedback
Arm D
Shuffled-feedback negative control

External outcome oracles

  • • Held-out target-state fidelity or independently sampled Hamiltonian energy — never the training reward
  • • Versioned manifest of target ID, split, policy hash, noise seed, shot budget, and resource logs
  • • Resource-matched negative-control comparison against shuffled feedback

Falsification conditions

  • • C fails to outperform the stronger of A and B on held-out outcomes.
  • • C fails to outperform the shuffled-feedback negative control D.
  • • An apparent advantage disappears under independent measurement seeds, calibrated noise, or matched resource accounting.