Nanchang Hangkong University researcher in continual learning for LLM fine-tuning and medical image segmentation.
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) 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—
Insufficient data: Evolve's snapshot for this researcher does not yet include a computable output score. This axis would measure publication volume and productivity over time, such as total papers and annual output trends.
Impact—
No impact score is available yet. This axis normally aggregates citation counts, h-index, and downstream uptake; both cataloged papers are recent 2026 arXiv preprints, so citation evidence has not accumulated and none is estimated here.
Focus0/100
The focus score is 0, reflecting an even 50/50 split in the field distribution between computer vision / medical image segmentation and machine learning (LLM fine-tuning and continual learning). The two cataloged papers map one-to-one onto these strands — SAUF-Net on segmentation, the post-hoc JANUS rectification work on continual learning — indicating a still two-sided rather than consolidated research agenda.
Collaboration64/100
Collaboration scores 64, based on 24 distinct coauthors in the bibliographic snapshot, with twelve recurring collaborators including Qin Lu, Zheyang Jing, Jianwang Li, Chen Yi, and Shaofeng Jiang. On both cataloged papers Yang appears as a middle-position author rather than first or corresponding author, suggesting contribution within team projects rather than independent group leadership so far.
Activity—
Activity data is not available: the snapshot lacks the per-year publication counts needed to score recent productivity. This axis would measure output cadence — papers per year and its trend — beyond the two 2026 arXiv preprints already reviewed.
Reproducibility55/100
Reproducibility averages 55 across the two reviewed papers (reproPapers: 2), both rated medium repro-readiness. Evolve's reviews flag concrete limitations: the continual-learning paper's replay set overlaps its evaluation set and its testing scope is narrow, while the segmentation work is evaluated only on 2D binary datasets — limitations addressable with broader, cleaner benchmarks.
Research summary
Yujie Yang is a researcher at Nanchang Hangkong University whose cataloged work to date spans two strands in roughly equal measure: semi-supervised medical image segmentation and machine learning methods for LLM fine-tuning and continual learning. Evolve's bibliographic snapshot lists two 2026 arXiv preprints, on both of which Yang is a middle-listed author rather than first or corresponding author. The first, SAUF-Net, proposes structure–appearance representation learning with uncertainty feedback for semi-supervised medical image segmentation; Evolve's review finds gains over AdaptFRCNet modest and evaluation limited to 2D binary datasets. The second applies post-hoc JANUS rectification to the stability–plasticity dilemma in fine-tuned LLMs via Jacobian null-space projection; the first-order theory is judged reasonable, but the testing scope is narrow and the replay set overlaps the evaluation set. Both papers carry a medium repro-readiness rating, averaging a reproducibility score of 55 across the two reviewed papers. Collaboration data records 24 distinct coauthors and a collaboration score of 64, with frequent collaborators including Qin Lu, Zheyang Jing, and Jianwang Li. Because "Yujie Yang" is a common name and no ORCID is provided, this profile is name-resolved from public bibliographic data only.
Collaboration pattern
Yang's recorded network spans 24 distinct coauthors, with a collaboration score of 64. Twelve collaborators recur, including Qin Lu, Zheyang Jing, Jianwang Li, Chen Yi, Shaofeng Jiang, Zhilong Zheng, Letian Tao, Yang Guan, Wei Xiong, Kehua Sheng, Bo Zhang, and Jingliang Duan. On both cataloged 2026 preprints Yang is a middle-position, non-corresponding author, consistent with contributing to team-led projects across both the segmentation and continual-learning strands.
Research keywords
Highlights
SAUF-Net: uncertainty feedback for semi-supervised medical image segmentation
Post-hoc Jacobian null-space rectification against forgetting in LLM fine-tuning
A two-track portfolio split evenly across fields
The bibliographic field distribution splits 50/50 between medical image segmentation and LLM fine-tuning/continual learning, and the two reviewed 2026 preprints each anchor one track. Both are method papers with medium repro-readiness, and Yang is a middle-listed, non-corresponding author on each.
Timeline
- 2026SAUF-Net preprint posted on arXiv: structure–appearance representation learning with uncertainty feedback for semi-supervised medical image segmentation; Evolve's review notes modest gains over AdaptFRCNet and 2D binary-only evaluation.
- 2026Post-hoc JANUS rectification preprint posted: Jacobian null-space projection to mitigate forgetting in LLM fine-tuning; Evolve notes a narrow testing scope and replay/evaluation set overlap.
- 2026Bibliographic snapshot records 24 distinct coauthors, an even 50/50 field split across two strands, and medium repro-readiness on both reviewed papers.
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.
- Dimension scores derive from a public bibliographic snapshot (OpenAlex-sourced); citation, h-index, and venue-normalized metrics were not available.
- The reviewed catalog contains only two papers — both 2026 arXiv preprints with middle-listed, non-corresponding authorship — so any wider publication record is not reflected here.
- "Yujie Yang" is a common name and no ORCID is provided; this profile is name-resolved from public bibliographic data, so identity disambiguation is uncertain.
- Output, impact, and activity axes are unscored due to insufficient data; no values are estimated for them.