Shun Zhang
China People's Public Security University · China
Speech/audio ML researcher building audio-language-model benchmarks and adaptation methods for telecom fraud detection
3
papers here
3
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-15) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
Output cannot be scored yet: the underlying publication-count and productivity data for this profile is insufficient. This axis would measure total research outputs and their distribution over time and across venues once more records are linked.
Impact—
Impact has no usable data yet — no citation counts or h-index are on file. This axis would measure citation-based scholarly influence, such as citations per paper and aggregate attention, once sufficient records exist.
Focus0/100
The radar computes focus at 0, driven by the raw field distribution: coverage splits evenly at 0.5/0.5 between two labels — audio-language model benchmarking for fraud detection, and audio-language models for fraud detection. In substance the two labels describe the same narrow applied line, so the low score reflects label fragmentation in the data rather than visible topic-hopping across the two 2026 catalog papers.
Collaboration66/100
Collaboration scores 58, built from 17 distinct coauthors across the two catalog papers, 12 of whom recur — including Huiyuan Liu, Qifan Wang, and Yifan Wang. The pattern points to stable team-based work rather than solo authorship, while middle-author positions on both papers and the absence of a corresponding role temper the score.
Activity40/100
Activity is not yet computable: the data lacks the continuous publication record needed to score it. This axis would measure recent publishing cadence and research continuity across years, rather than isolated catalog entries.
Reproducibility85/100
Reproducibility is the strongest recorded axis at 78, a mean computed over 2 papers. Evolve rates TeleAntiFraud 2.0's reproducibility readiness high and FRAUDSkill's medium, so the score rests on benchmark artifacts and documented ablations, tempered by synthetic-only data and thin human validation on the benchmark side.
Research summary
Shun Zhang is an early-stage researcher in speech and audio machine learning whose visible work centers on audio-language models for telecom fraud detection. The Evolve catalog holds two 2026 arXiv papers, both with middle-author positions and no corresponding role recorded. The first, TeleAntiFraud 2.0, is a refreshable, profile-grounded, audio-based benchmark for telecom fraud detection; Evolve's review credits its near-domain negative design and collapse-aware analysis while flagging synthetic-only data and thin human validation, and rates its reproducibility readiness high. The second, FRAUDSkill, applies structured frozen-weight skill optimization to audio anti-fraud detection, reporting a +31.96 Macro-F1 gain on the TeleAntiFraud benchmark; Evolve's ablation reading attributes most of that gain to structured post-processing rather than the skill search itself, and rates reproducibility readiness medium. Together the two papers form a benchmark-plus-method pairing around one applied problem. The bibliometric radar scores reproducibility at 78 (a mean over 2 papers) and collaboration at 58 from 17 distinct coauthors, while the focus score sits at 0 under an even 0.5/0.5 split between two closely related field labels. Output, impact, and activity metrics are not yet available for this profile.
Collaboration pattern
Across the two catalog papers Shun Zhang records 17 distinct coauthors, with 12 recurring collaborators including Huiyuan Liu, Qifan Wang, Yifan Wang, Yijin Zhou, and Yuxi Zhao — a dense pattern consistent with team-based AI research. Both papers carry middle-author positions with no corresponding role, suggesting a contributor rather than lead within the group, in line with the radar collaboration score of 58.
Research keywords
Highlights
TeleAntiFraud 2.0: a refreshable audio benchmark for telecom fraud detection
FRAUDSkill: frozen-weight skill optimization reporting +31.96 Macro-F1
Shortcut learning and prediction collapse treated as first-class problems
Timeline
- 2026Publishes TeleAntiFraud 2.0, a refreshable, profile-grounded, audio-based benchmark for telecom fraud detection (arXiv, middle-author position).
- 2026Publishes FRAUDSkill, applying structured frozen-weight skill optimization to audio anti-fraud detection on the TeleAntiFraud benchmark (arXiv, middle-author position).
- 2026Both papers enter the Evolve catalog as a benchmark-plus-method pairing, with reproducibility readiness rated high and medium respectively.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric scores are computed from OpenAlex-based public bibliographic data only; no institutional affiliation, country, or ORCID is on file.
- “Shun Zhang” is a common name; this profile is name-resolved from public bibliographic records and may include or omit namesake work.
- Output, impact, and activity axes are marked data-insufficient, so no publication counts, citation totals, or productivity trends are quoted anywhere in this profile.
- Only two 2026 arXiv preprints are in scope, both middle-author and non-corresponding; no peer-reviewed venue record is present.