Early-stage researcher building speech and multimodal video benchmarks with strong reproducibility signals.
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
2
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—
Not yet scored: full publication-volume data is unavailable for this identity. This axis would measure total output and career-level productivity from a complete publication record, beyond the two papers in the Evolve catalog and the sparse OpenAlex year counts currently on file.
Impact—
Not yet scored: citation-impact data is insufficient. Beyond the recorded h-index of 2 and i10 index of 1, no citation counts or uptake metrics are available; this axis would measure citations, scholarly influence, and downstream use of the benchmark resources.
Focus0/100
Scores 0 because output splits evenly across two distinct areas: Evolve's field distribution assigns 0.5 to computational linguistics/speech technology and 0.5 to computer vision/multimodal machine learning. OpenAlex topics echo the spread — speech recognition and synthesis (3), intellectual property and patents (2), and topic modeling, scientometrics, and NLP techniques (1 each). The work is linked by benchmark construction, but no single field yet dominates.
Collaboration50/100
Scores 50, reflecting a moderate collaborative footprint: 11 distinct coauthors appear in the record, and both catalogued 2026 papers are middle-author, multi-person efforts rather than solo work. The frequent-coauthor roster — including Seoyeon An, Hyeonseo Jang, Minsu Kim, Kangwook Lee, Jong-Hwan Kim, and Dong-Jin Kim — points to a concentrated, largely Korea-linked network built on repeated collaboration rather than wide cross-institutional spread.
Activity47/100
Scores 47, consistent with an emerging or recently reactivated record. OpenAlex counts by year show 1 work in 2025 and 3 in 2026 — including both Evolve-catalogued arXiv preprints — with 4 works over the last three years. An isolated 2005 cluster of 2 works sits far outside this active window and likely reflects name ambiguity rather than current output.
Reproducibility78/100
Scores 78, the strongest measured axis, computed as the mean reproducibility rating across 2 reviewed papers. AgentVidBench is rated high readiness — a released 100-question benchmark with LLM-judged trajectory scoring and validated judge behavior — while the phonemization benchmark is rated medium, with Evolve noting that its Korean gold data is noisy. Benchmark-and-resource outputs inherently favor external reuse.
Research summary
Younghan Park, affiliated in Evolve's data with Carnegie Mellon University and name-resolved from OpenAlex records, is an early-stage researcher whose catalogued work centers on evaluation infrastructure for language and multimodal systems. The Evolve catalog holds two 2026 arXiv preprints, both with Park in a middle author position: a multilingual grapheme-to-phoneme benchmark for phonemizing user-generated text, and AgentVidBench, a multi-hop video question answering benchmark for evaluating MLLM agents. Evolve's field distribution splits evenly (0.5/0.5) between computational linguistics/speech technology and computer vision/multimodal machine learning, echoed by OpenAlex topics led by speech recognition and synthesis (3 works), alongside intellectual property and patents (2), topic modeling, scientometrics, and NLP techniques (1 each). Bibliometric signals are thin: h-index 2, i10 index 1, 11 distinct coauthors, 4 works in the last three years, with indexed counts of 1 work in 2025 and 3 in 2026, plus an isolated 2005 cluster likely reflecting name ambiguity. Reproducibility is the strongest measured axis, averaging 78 across the two reviewed papers — consistent with benchmark-and-resource output designed for external reuse.
Collaboration pattern
Park works in multi-author teams, holding middle author positions on both 2026 catalog papers, neither as corresponding author. The record lists 11 distinct coauthors, with a frequent-coauthor roster including Seoyeon An, Hyeonseo Jang, Minsu Kim, Chanho Lee, Kangwook Lee, MinJu Jeon, Han Sung Park, Jong-Hwan Kim, Dong-Jin Kim, and Hoyeon Lee — a largely Korea-linked network with repeated collaboration. The collaboration score of 50 reflects moderate breadth without evidence of wide international spread.
Research keywords
Highlights
AgentVidBench: multi-hop video QA benchmark for MLLM agents
First multilingual benchmark for phonemizing user-generated text
Reproducibility-first benchmark building across two fields
Across an even 0.5/0.5 split between speech technology and multimodal video understanding, both catalogued papers are evaluation resources, yielding a mean reproducibility score of 78 — the highest measured dimension in this profile.
Timeline
- 2005OpenAlex indexes 2 works to this identity in 2005 — far predating the current record and likely a namesake cluster; treated with caution in this profile.
- 2025OpenAlex indexes 1 work in 2025, marking the start of the current active window that spans 4 works over the last three years.
- 2026Most active indexed year with 3 works, including both Evolve-catalogued arXiv preprints — the multilingual UGT phonemization benchmark and AgentVidBench — both with Park in a middle author position.
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 data comes from OpenAlex only; identity is name-resolved without ORCID, so same-name authors may be mixed in — the 2005 cluster and the patent/scientometric topics may belong to a different Younghan Park.
- Only 2 papers have been reviewed by Evolve; the output and impact dimensions are unscored due to insufficient data.
- Both catalogued works are 2026 arXiv preprints with middle authorship; peer-review status is not reflected in this data.
- h-index 2 and i10 index 1 describe a very early-stage record; scores are volatile and will shift quickly with new publications.