Shenzhen University researcher bridging learnable sparse coding and multimodal LLM hallucination analysis
2
papers here
2
reviewed by Evolve
2
first-author
0
corresponding
—
works · OpenAlex
1
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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—
Unscored: OpenAlex coverage of this early-stage record is too thin to rate publication volume (h-index 1, i10-index 0, 6 recent three-year works). This axis would measure total research output and its growth across years and publication venues.
Impact—
Unscored: the citation footprint is essentially absent in the snapshot (h-index 1, i10-index 0), partly because both reviewed papers are 2026 arXiv preprints. This axis would measure citations, usage, and downstream scholarly influence accumulating over time.
Focus0/100
The focus score is 0: the OpenAlex field distribution splits exactly 50/50 between computer vision / machine learning and multimodal machine learning, and five topic clusters coexist — Advanced Neural Network Applications (5), Sparse and Compressive Sensing Techniques (2), plus Topic Modeling, Natural Language Processing Techniques, and Computational Drug Discovery Methods (1 each). This breadth reads as a deliberately exploratory, method-driven agenda rather than a single niche.
Collaboration41/100
Scoring 41, collaboration is moderate for an early-career record: 7 distinct coauthors appear across the recent three-year window, with recurring partners including Youlu Xing, Yinchen Liu, Mingxuan Cui, Junye Chen, Jucheng Liu, Song Wang, and Ruize Han. Both Evolve-reviewed papers are first-authored but not corresponding-author work, consistent with contributing inside a small team rather than leading a large consortium.
Activity57/100
Activity scores 57, backed by 6 works in the recent three-year window and a clear acceleration in annual counts: 1 work in 2023, 2 in 2024, and 4 in 2026 — the 2026 group includes both arXiv preprints reviewed by Evolve. The cadence points to an actively producing phase, though the record is too short to establish a long-term trend.
Reproducibility55/100
Reproducibility averages 55 across the 2 papers Evolve has assessed, both rated medium readiness. The reviews temper reported gains with concrete caveats: the sparse coding paper was evaluated in a single run under noise-only settings, while HEAL needs per-benchmark tuning and offers thin causal evidence. The methods described (FISTA unrolling, inference-time calibration) appear amenable to independent reimplementation.
Research summary
Meng'en Qin is an early-stage researcher affiliated with Shenzhen University; this profile is name-resolved from public OpenAlex bibliographic data. The record splits evenly between computer vision / machine learning and multimodal machine learning, and topic clustering is led by Advanced Neural Network Applications (5 works), with Sparse and Compressive Sensing Techniques (2) plus Topic Modeling, Natural Language Processing Techniques, and Computational Drug Discovery Methods (1 each). Two 2026 first-authored arXiv preprints have been reviewed by Evolve: an adaptive convolutional sparse coding method whose sparsity coefficient is learned through FISTA unrolling with label-free λ adaptation, and HEAL, which attributes multimodal LLM hallucination to information distribution drift in synergy heads and applies inference-time calibration. Evolve credits strong reported gains in both, while flagging single-run, noise-only evaluation in the first and per-benchmark tuning with thin causal evidence in the second; both carry medium reproducibility readiness. Bibliometric signals remain thin — h-index 1, i10-index 0, 6 works in the recent three-year window, 7 distinct coauthors, and annual counts of 1 (2023), 2 (2024), and 4 (2026) — so the output and impact axes are left unscored. The overall picture is an exploratory, method-driven early career whose citation footprint has yet to form.
Collaboration pattern
With 7 distinct coauthors over the recent three-year window and a collaboration score of 41, Meng'en Qin works in small teams rather than large consortia. Recurring partners include Youlu Xing, Yinchen Liu, Mingxuan Cui, Junye Chen, Jucheng Liu, Song Wang, and Ruize Han. Both Evolve-reviewed papers are first-authored but not corresponding-author work, suggesting an emerging contributor building joint projects across computer vision and multimodal machine learning.
Research keywords
Highlights
Learnable sparsity for convolutional sparse coding
HEAL: tracing MLLM hallucination to synergy-head drift
Timeline
- 2023Earliest indexed output in the OpenAlex snapshot: 1 publication in this year.
- 2024Indexed output doubles to 2 works, maintaining an early, low-volume cadence.
- 2026First-authored arXiv preprint on adaptive convolutional sparse coding via information bottleneck is reviewed by Evolve (medium reproducibility readiness).
- 2026Second first-authored preprint, HEAL on synergy-head information drift in MLLM hallucination, enters the Evolve catalog; indexed works total 4 in this year.
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 indicators come from an OpenAlex snapshot only; no ORCID is linked to this profile.
- Both reviewed papers are 2026 arXiv preprints; peer-review status and venue acceptance are not reflected in the data.
- Output and impact are unscored because citation coverage is too sparse at snapshot time (h-index 1, i10-index 0).
- Identity is name-resolved from public bibliographic data; similarly named researchers may be merged or split in the underlying record.