National Taiwan University NLP researcher working on large language model adaptation and biomedical text simplification
2
papers here
2
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 data is insufficient to score this axis. The snapshot records only 12 works in 2026; a full output measure would need the complete publication count, venue mix and yearly distribution to gauge productivity.
Impact—
Impact data is insufficient to score this axis. The snapshot's h-index and i10-index read zero, so citation-based influence cannot be assessed; this axis would measure citations, uptake and scholarly reach of the work.
Focus0/100
The focus score is 0/100. The snapshot's field distribution is an even 50/50 split between biomedical text simplification and LLM continual adaptation, and its topics span five clusters — topic modeling (12), NLP techniques (11), domain adaptation and few-shot learning (5), multimodal applications (5) and speech (1) — so no single line dominates this early snapshot.
Collaboration39/100
The collaboration dimension scores 39/100 from 6 distinct coauthors in the snapshot. The frequent-coauthor list repeats Ting-Wei Chang, Hen-Hsen Huang and Po-Chun Chen, and both reviewed 2026 papers are last-authored, non-corresponding positions, pointing to a compact recurring team rather than a wide collaborative network.
Activity75/100
The activity dimension scores 75/100. The snapshot's counts-by-year attribute all 12 recorded works to 2026, and the recent three-year work count is also 12, meaning the entire captured output falls in a single recent year — an active but short-window record rather than a long publication history.
Reproducibility38/100
The reproducibility dimension averages 38/100 over the 2 papers assessed. The medical text simplification paper is rated low — Evolve notes unsupported mixture-of-agents gains and absent statistics — while the continual LLM adaptation paper is rated medium, with real but task-dependent gains evaluated on a single seed.
Research summary
Hsin-Hsi Chen is a researcher at National Taiwan University whose captured work sits at the intersection of large language model adaptation and biomedical text simplification. The bibliometric snapshot, resolved from OpenAlex, records 12 works in 2026, matching the recent three-year window, with an even 50/50 split across the two fields. Both reviewed papers are 2026 arXiv preprints carrying Chen as last author, consistent with a senior or supervisory role rather than corresponding authorship. The first presents a broad PLABA-2024 study of prompting, QLoRA finetuning and mixture-of-agents for adapting medical texts into plain language; Evolve credits a solid empirical sweep but finds the mixture-of-agents gains unsupported and statistical testing absent. The second proposes environment-driven dynamic policies for continual LLM improvement, with gains judged real but task-dependent and evaluation limited to one seed. Topic signals span topic modeling, core NLP techniques, domain adaptation and few-shot learning, multimodal applications and speech, producing a focus score of 0. Collaboration rests on 6 distinct coauthors including Ting-Wei Chang, Hen-Hsen Huang and Po-Chun Chen, scoring 39, while activity scores 75 on the concentrated 2026 output. Reproducibility averages 38 across the two papers. Citation metrics read zero in the snapshot and the output and impact axes remain unscored, so this should be read as an early-stage, name-resolved profile.
Collaboration pattern
The snapshot resolves 6 distinct coauthors, and Evolve's frequent-coauthor list centers on Ting-Wei Chang, Hen-Hsen Huang and Po-Chun Chen. Both reviewed 2026 papers carry Chen as last, non-corresponding author, suggesting a senior collaborator pattern built on recurring partners rather than a large distributed network; the collaboration dimension scores 39 on this basis.
Research keywords
Highlights
Plain language adaptation of medical texts at PLABA-2024 scale
Environment-driven dynamic policies for continual LLM improvement
Senior last-author role across the 2026 output
On both reviewed 2026 papers Chen is the last and non-corresponding author, spanning biomedical text simplification and LLM continual adaptation — the two fields that split the snapshot's field distribution 50/50.
Timeline
- 2026Snapshot records 12 works, all falling in 2026 and matching the recent three-year count.
- 2026Last-author arXiv study on LLM-driven plain language adaptation of medical texts (PLABA-2024).
- 2026Last-author arXiv paper proposing environment-driven dynamic policies for continual LLM improvement.
- 2026Reproducibility assessed on the two reviewed papers with a mean score of 38 (one low, one medium).
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 are computed from an OpenAlex snapshot only; no ORCID was available for identity confirmation, so this profile is name-resolved from public bibliographic data.
- Citation-level metrics (h-index, i10-index) read zero in the snapshot and the output and impact radar axes are unscored; these figures likely reflect incomplete author disambiguation rather than measured impact.
- Only two papers from the Evolve catalog underlie the reproducibility and positioning assessments; coverage of the researcher's full publication record is not guaranteed.
- All catalog papers are 2026 arXiv preprints; peer-reviewed venue status and career facts fall outside the provided data.