Chanyoung Park
Korea Advanced Institute of Science and Technology
KAIST researcher linking graph-based corporate default prediction with LLM-based user simulation in recommender systems
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-08) 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: the catalog holds only two 2026 arXiv preprints, too thin to measure publication output. This axis would total peer-reviewed papers, weigh venue quality, and track output cadence across the full publication record.
Impact—
Not yet scored: no citation or influence data is available in the provided snapshot. This axis would measure citations, downstream adoption, and scholarly attention to the two preprints once citation indexes cover them.
Focus0/100
Scores 0/100: the reviewed record splits evenly — 0.5 machine learning for finance (credit risk) and 0.5 recommender systems / LLM-based user simulation — with no dominant field. Both 2026 papers are last-authored in these separate lines, showing parallel work in graph-based credit modelling and LLM-driven evaluation rather than one convergent agenda.
Collaboration46/100
Scores 46/100: the two reviewed papers share 9 distinct frequent coauthors — including Junghoon Kim, Hyunsung Kim, and Seungyoon Choi — with neither paper listing this researcher as corresponding author; both are last-authored, indicating senior involvement in team-based projects rather than solo or corresponding-lead work.
Activity—
Not yet scored: with only two cataloged 2026 preprints, yearly output patterns cannot be established. This axis would track publication frequency and sustained research pace over multiple years of the record.
Reproducibility38/100
Scores 38/100, the mean across 2 assessed papers. The LLM user-simulator study carries medium reproducibility readiness, though hand-tuned thresholds raise questions; DefaultGNN scores low because its large Korean tax-invoice evaluation rests on proprietary transaction data external researchers cannot access.
Research summary
Chanyoung Park is a researcher at the Korea Advanced Institute of Science and Technology (KAIST) whose reviewed work spans two distinct lines: machine learning for finance (credit risk) and recommender systems with LLM-based user simulation. Evolve's catalog holds two 2026 arXiv preprints, both with Park in the last author position and neither as corresponding author. The first, DefaultGNN, proposes a dual-perspective graph neural network that reads buyer-seller transaction networks from both buyer and seller views to predict corporate default; Evolve's review calls the large Korean tax-invoice study solid, while noting that proprietary data access and modest gains temper its reach. The second diagnoses trait interference in LLM user simulators, showing how a behavioral trait leaks into preference scores; the review finds the diagnosis convincing and the prompt-based PQA intervention empirically supported, with hand-tuned thresholds leaving open questions. The bibliometric record splits evenly between the two fields (0.5 and 0.5), and the papers share a pool of 9 distinct frequent coauthors. Reproducibility readiness averages 38 across the two assessed papers, limited mainly by proprietary evaluation data. With citation, output, and activity measures not yet computable from the available snapshot, this remains an early-stage profile whose full record is likely broader than the cataloged preprints.
Collaboration pattern
Both cataloged 2026 papers are last-authored, non-corresponding works built with a pool of 9 distinct frequent coauthors — Junghoon Kim, Hyunsung Kim, Seungyoon Choi, KyoungYong Park, Jihun Lee, YongGu Ji, Chaehyun Kim, Sein Kim, and Hongseok Kang. The pattern indicates senior, last-author involvement in collaborative team projects across both research lines, rather than solo writing or corresponding-author leadership.
Research keywords
Highlights
Dual-perspective GNN for corporate default prediction
When behavioral traits leak into simulated preferences
One senior-author seat, two evaluation problems
Across both 2026 preprints Park sits in the last author position, connecting two communities: graph-based credit-risk learning on transaction networks and LLM-based user simulation for recommender evaluation, with the field split at 0.5 and 0.5.
Timeline
- 2026Last-authored DefaultGNN preprint proposes a dual-perspective GNN for corporate default prediction from buyer-seller transaction networks (arXiv).
- 2026Last-authored study diagnoses trait interference in LLM user simulators, showing a behavioral trait leaking into simulated preferences (arXiv).
- 2026Both cataloged outputs land in the same year, splitting evenly between credit-risk graph learning and LLM-based user simulation, with 9 distinct frequent coauthors across the record.
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.
- Radar scores are derived from public bibliometric data (OpenAlex-based); output, impact, and activity lacked sufficient data in this snapshot.
- The Evolve catalog holds only two 2026 arXiv preprints for this researcher; the profile is early-stage and likely under-represents the full publication record.
- 'Chanyoung Park' is a common Korean name; this profile is name-resolved from public bibliographic data and the listed institution, with no ORCID verification.
- Frequent coauthors come from the provided co-author list; per-paper author attribution was not available in this snapshot.