KJ

Kyomin Jung

Seoul National University · South Korea

Unclaimed · AI-structured profileORCID OpenAlex

Seoul National University researcher in conversational AI, efficient LLM inference, and information retrieval.

3

papers here

3

reviewed by Evolve

0

first-author

0

corresponding

—

works · OpenAlex

27

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: insufficient dataOutputImpact: insufficient dataImpactFocus: 0/100FocusCollaboration: 51/100CollaborationActivity: 100/100ActivityReproducibility: 55/100Reproducibility

Output—

Total publication-volume metrics are not yet available in this snapshot, so raw output cannot be scored. This axis would measure cumulative works, annual production totals, and the output trajectory relative to field norms.

Impact—

Citation counts, citation percentiles, and other influence measures are absent from the current data, so impact is not scored. This axis would capture aggregate citations, h-index-based influence, and the reception of individual works.

Focus0/100

The focus score is 0, consistent with the recorded field distribution: natural language processing (conversational AI), machine learning (efficient LLM inference), and NLP/information retrieval each hold exactly one third, with no dominant line. Topic clusters likewise spread across Topic Modeling (88), NLP Techniques (60), graph algorithms (26), multimodal ML (23), and neural networks (18).

Collaboration51/100

The collaboration score is 51, reflecting a compact but stable network: 12 distinct recent coauthors appear in the snapshot, and 12 frequent collaborators are listed, including Hyojeong Yu, Hyukhun Koh, Sunghyeon Woo, and Jahyun Koo. The three 2026 preprints place Kyomin Jung in middle or last author positions without a corresponding role, suggesting contributions embedded in a small, recurring team rather than a large distributed consortium.

Activity100/100

Activity is scored at the maximum 100. The trailing three years contain 61 works, and the annual series shows sustained growth from single digits in the 2000s to 16 in 2019, 23 in 2021, and a plateau of 19, 21, and 21 across 2024–2026 — an unusually high and steady recent tempo.

Reproducibility55/100

Reproducibility scores 55, computed as a mean readiness of 55 across the 3 Evolve-reviewed papers. Each preprint is rated medium readiness: Carryover Drafting carries epoch-fairness and no-code caveats, CompCQR hinges on rank-1 picking with single-run evaluations, and PRAGMA is solid but synthetic, small-scale, and judge-circular.

Research summary

Kyomin Jung (Seoul National University, South Korea) is a name-resolved profile built from OpenAlex bibliographic records. The indexed record spans more than two decades, with an h-index of 27 and an i10-index of 64. Topical clustering places the largest share of work in Topic Modeling (88) and Natural Language Processing Techniques (60), with smaller clusters in Complexity and Algorithms in Graphs (26), Multimodal Machine Learning Applications (23), and Advanced Neural Network Applications (18). Recent activity is high: 61 works in the trailing three years, with annual counts peaking at 23 in 2021 and remaining elevated at 19, 21, and 21 across 2024–2026. Current field distribution is split evenly across three areas — natural language processing (conversational AI), machine learning (efficient LLM inference), and NLP/information retrieval — each at exactly one third, which is why no dominant research line emerges. The Evolve-reviewed catalog contains three 2026 arXiv preprints: Carryover Drafting on recycling rejected verifier states for speculative decoding, CompCQR on training-free conversational query reformulation, and PRAGMA, a 400-query benchmark for personalized guidance with long-term conversational memory. In all three the author holds a middle or last position rather than a corresponding role. Collaboration breadth in this snapshot is compact, with 12 distinct recent coauthors and a recurring group including Hyojeong Yu, Hyukhun Koh, and Sunghyeon Woo.

Collaboration pattern

The collaboration network is compact: 12 distinct recent coauthors in the snapshot and 12 frequent collaborators, including Yunah Jang, Hyojeong Yu, Hyukhun Koh, Minsung Kim, and Jahyun Koo. On the three 2026 preprints Kyomin Jung appears in middle or last author positions, none as corresponding author, consistent with contributing within a stable recurring group rather than leading large multi-institution consortia.

Research keywords

speculative decodingLLM inference accelerationKV cache reuseconversational searchconversational query reformulation (CQR)dense retrievallong-term conversational memorypersonalized guidanceretrieval-augmented generationtopic modelingLLM benchmarkingtraining-free methods

Highlights

Timeline

  1. 2005First year in the indexed record, with 1 work in the annual counts.
  2. 2019Output reaches 16 works, the highest annual count of the 2010s.
  3. 2021The record's peak annual count: 23 works.
  4. 2024Sustained high output begins, with 19 works this year.
  5. 2025Output rises to 21 works, continuing the elevated pace.
  6. 2026Three arXiv preprints reviewed in the Evolve catalog (Carryover Drafting, CompCQR, PRAGMA); the annual count reaches 21.

Papers in the Evolve catalog

Frequent co-authors

Is this you? Claim this profile

Enter the email you published with. We'll send a one-time claim link; after claiming you can edit the introduction or hide this page. Requests from a mismatched address go to manual review.

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 (h-index, i10-index, counts by year, topics) are derived from OpenAlex records and reflect only that source's indexing.
  • The output and impact dimensions are unscored because total publication counts and citation metrics were not provided in this snapshot.
  • Evolve's individual review covers 3 arXiv preprints from 2026; the author's other works are not individually assessed.
  • The profile is name-resolved from public bibliographic data (OpenAlex identity) and does not incorporate verified biographical or career information.
Kyomin Jung — Researcher Profile | Evolve.Pub