Geng Academic Fraud Investigation Report: "Scaling crossovers in non-equilibrium critical dynamics" (Li et al. 2025, DOI: 10.1088/1751-8121/ae066a)

Academic fraud report · Geng Detector

Summary

This investigation report examines the paper "Scaling crossovers in non-equilibrium critical dynamics" by Rong Li, Qirui Ding, and Weicheng Cui, published in J. Phys. A: Math. Theor. 58 (2025) 385004. The verdict is that the paper is CLEAR (no academic misconduct detected). Three independent checks were performed: (1) originality of figures and data confirmed, with all validation data drawn from published literature (Takeuchi et al. 2007/2009 liquid crystal data and Allam et al. 2013 carbon nanotube data) and R^2 values exceeding 0.998 consistent with theoretical scaling-law fits; (2) timeline and AI-disclosure compliance verified, with submission on 2025-04-26, publication on 2025-09-23, and explicit acknowledgment of O3-mini-high language polishing; (3) mathematical self-consistency of reported R^2 and RMSE values (e.g., 1.7±0.2, 0.45±0.1) confirmed as natural nonlinear least-squares fit outputs across models N=2 to N=6. No further action is recommended.

Verdict

Cleared — No academic misconduct detected. The six-form screening procedure returned no anomalies. The paper is a purely theoretical physics and mathematical modeling study with no fabricated experimental images or data.

Key findings

Figure and data originality (Forms 1 & 2): All validation data are sourced from previously published literature, including liquid crystal experiments by Takeuchi et al. (2007/2009, refs [18,19]) and carbon nanotube exciton data by Allam et al. (2013, ref [46]). Fit quality R^2 ≥ 0.998 is consistent with theoretical scaling-law expectations.
Timeline and AI-use compliance (Forms 5 & 6): Submission on 2025-04-26 and publication on 2025-09-23 are internally consistent. The authors transparently disclose AI language polishing with O3-mini-high and accept full responsibility for content — fully compliant with current publication ethics.
Statistical self-consistency (Forms 3 & 4): Reported R^2 and RMSE values in Table A1 (e.g., 1.7 ± 0.2, 0.45 ± 0.1) display natural variation characteristic of nonlinear least-squares fitting across models N=2 through N=6. No signs of artificially uniform random-number generation.
Open data/code availability: Authors have publicly released code and data on GitHub.

Evidence highlights

DOI: 10.1088/1751-8121/ae066a
Journal: J. Phys. A: Math. Theor. 58 (2025) 385004
Reported fits: R^2 ≥ 0.998 across all datasets
Representative error ranges from Table A1: 1.7 ± 0.2; 0.45 ± 0.1
AI disclosure quote: "The language of the manuscript has been polished using the O3-mini-high model."
Timeline: submitted 2025-04-26; published 2025-09-23

Notes

This is a theoretical paper; therefore image-based forgery checks (Western blot, flow cytometry) are inapplicable.
The AI-polishing disclosure is exemplary and well above current minimum disclosure norms.
The report explicitly acknowledges that final determination of misconduct requires institutional investigation and that false positives/negatives are possible.