Geng Academic Fraud Detection Report: 'Strategy-Hiding Bilateral Attribute-Based Access Control Scheme'

Academic fraud report · Geng Detector

Summary

Verdict: Questionable (🟡). This report, covering DOI 10.13451/j.sxu.ns.2025008, focuses on a cryptographic access control paper whose core mathematical derivations (bilinear pairings in Section 4.1–4.2) appear internally consistent and free of algebraic fabrication. However, serious production-quality and editorial-integrity issues undermine confidence. The Figure 2 caption contains obvious square-bracket placeholders (e.g., '[Python 3.8.0 with Charm 0.43 library and MNT224 elliptic curve environment]'), strongly suggesting unreviewed AI-assisted translation or templating residue. Reference [13] lists 'RIEPEL D, WEE H, RIEPEL D, et al.', repeating the first author and misusing 'et al.' for a two-author entry—a typical metadata/AI-hallucination artifact. No image manipulation or numerical falsification could be confirmed because experiment figures were not textually verifiable. Recommended actions are editorial corrections rather than a misconduct referral.

Verdict

Questionable (🟡). The theoretical core passes spot-checks, but the manuscript exhibits clear AI-template residue and reference-list errors indicative of insufficient human proofreading and weak editorial scrutiny.

Key findings

AI placeholders left in Figure 2 caption: Both Chinese and English captions end with bracketed placeholder text: [Python 3.8.0搭配Charm 0.43库和MNT224椭圆曲线环境] and [Python 3.8.0 with Charm 0.43 library and MNT224 elliptic curve environment], with time unit also bracketed as [毫秒] / [milliseconds].
Reference [13] metadata corruption: Author list rendered as RIEPEL D, WEE H, RIEPEL D, et al., duplicating the first author and appending et al. to a two-author paper (Doreen Riepel and Hoeteck Wee).
Core cryptographic derivation verified (negative finding): Spot-checked bilinear-pairing equation in Section 4.2, e(A_1, B_2) = e(g_1, g_2)^{b_2 · a/b} = e(B_1, B_2)^α, is algebraically consistent under α = a/b; the Verify algorithm in Section 4.1.2 showed no logical break.
Performance discussion credible (negative finding): Authors acknowledge slightly higher encryption cost than the MABE baseline in Figure 2b due to policy hiding, and claim decryption under 1 ms—plausible and not in the 'too perfect' failure pattern.
No image-manipulation evidence: Figures 2 and 3 are described textually only; pixel-level forensics not performed.

Evidence highlights

DOI: 10.13451/j.sxu.ns.2025008
Journal: Journal of Shanxi University (Natural Science Edition), online first 2025-04-15.
Figure 2 caption literal excerpt (CN): 注：实验在 [Python 3.8.0 搭配 Charm 0.43 库和 MNT224 椭圆曲线环境] 下进行，时间单位为 [毫秒]。
Figure 2 caption literal excerpt (EN): Note: The experiments were conducted in [Python 3.8.0 with Charm 0.43 library and MNT224 elliptic curve environment], and the time is measured in [milliseconds].
Reference [13] literal excerpt: RIEPEL D, WEE H, RIEPEL D, et al.
Verified equation: e(A_1, B_2) = e(g_1, g_2)^{b_2 · a/b} = e(B_1, B_2)^α.

Notes

All findings reflect production-quality defects (AI residue, bibliographic errors), not proven research misconduct; theoretical content is internally consistent under the limited checks performed.
Limitations: no pixel-level image analysis; bibliographic review limited to the single flagged reference; AI-residue inference is based on formatting heuristics and is not definitive proof of AI authorship.
Recommended follow-up: contact authors and editorial office to correct caption placeholders and reference [13]; consider posting on PubPeer. The report does not recommend a formal misconduct referral at this time.