Works
4
Sources
13
Records
35,036
Period
900–1500
Last revised
28 August 2026

samhan.ai

Computational Inquiry into Medieval and Early Modern Korea

A research site by Javier Cha

Samhan brings together critical editions, structured historical data, research applications, and computational experiments for the study of medieval and early modern Korea. These works explore forms of reading and historical inquiry made possible by digital environments, computation, and machine learning that take scholarship beyond the printed page.

Sebo世譜 세보BETA]

Id
SH/A1
Type
APPLICATION
Period
918–1600
Version
0.9

Medieval and Early Modern Korean Kinship Workbench. Node and edge tables mapped back to the passage each was read from. Every identification carries its agent and its date; corrections are appended to a ledger and retired identifiers are never reissued.

3 corpora · 13 sources2026-08 sebo.samhan.ai

The Andong Kwŏn Genealogy of 1476安東權氏成化譜 안동권씨 성화보]

Id
SH/D1
Type
DATA
Period
1476
Version
1.0

A digital critical edition. The oldest extant comprehensive Korean genealogy, edited from the Academy of Korean Studies facsimile, with an inspection apparatus that checks the structured data against the original leaf by leaf.

facsimile · node and edge tables · notebooks2025-06-20 andongkwon1476.samhan.ai

Koryŏ Epitaphs高麗墓誌銘 고려묘지명BETA]

Id
SH/D2
Type
DATA
Period
918–1392
Version
0.4

A structured corpus in five parallel layers. From the reading of the stone to the extracted relations, with a non-destructive overlay: no layer above may alter the layer beneath it, and the reading of the stone is never rewritten.

341 documents · 5 layers2026-08 epitaphs.samhan.ai

Closure通婚圈 통혼권ALPHA]

Id
SH/X1
Type
EXPERIMENT
Period
1300–1500
Version
0.2

A computational experiment on marriage-circle closure. Whether tightly closed in-group marriage gave way to more diffusely embedded ties, tested against the Sŏnghwabo edition. The first run did not support the hypothesis and is published as it stands.

notebook · figures · a negative result kept2026-07 closure.samhan.ai

Sources

The historical corpora these works are made from, and what each contributes. The meter is that source's share of all 35,036 records in Samhan — one tally per ten per cent — so the weight of the Sŏnghwabo against the rest is visible rather than asserted. 13 corpora are configured; the 6 that currently contribute records are below.

Andong Kwŏn Sŏnghwabo安東權氏成化譜
46%
7,496 nodes · 8,778 edges Academy of Korean Studies facsimile
SH/D1 SH/X1
Koryŏ epitaphs高麗墓誌銘
15%
2,871 nodes · 2,510 edges 341 documents in five layers
SH/D2 SH/A1
Koryŏsa, biographies高麗史 列傳
13%
2,717 nodes · 1,906 edges 1,451 biographies read
SH/A1
Examination rosters to 1494榜目
12%
3,540 nodes · 796 edges reconciled against the AKS database
SH/A1
Sillok obituaries實錄 卒記
8%
1,581 nodes · 1,252 edges in preparation
SH/A1
Koryŏsa Chŏryo高麗史節要
5%
778 nodes · 811 edges in preparation
SH/A1

Timeline

Milestones in digital Korean studies, 2008 to the present.

200820122016202020242026
2019Big Data Studies Lab foundedSeoul National University.
2020Digital Historian-in-ResidenceLingnan University, Hong Kong.
2022Lab relocated to Hong KongUniversity of Hong Kong.
2025Andong Kwŏn Genealogy of 1476First public release, 20 June.
2025Opening keynote, DH2025Lisbon.
2026Sebo enters betaSamhan.ai published.

Colophon

Javier Cha · Department of History, University of Hong Kong

Samhan.ai · Version 1.0 · Revised 28 August 2026

Text CC BY 4.0 · Data CC0 · Code MIT

Annual snapshot preserved in the HKU institutional repository.

Cite

Cha, Javier. Samhan.ai: Computational Inquiry into Medieval and Early Modern Korea. Version 1.0, 2026.