Multimodal contrastive learning traces the early dissemination of Romance of the Three Kingdoms in Thailand

Abstract The Romance of the Three Kingdoms (RTK, 三國演義) is a global literary canon, yet how and when it first reached Thailand remains contested because few documents survive. Using the 1838 Wat Prasert Sutthawat murals, the earliest large-scale Thai RTK visual corpus, we build the first expert-verified multimodal dataset for mural–print comparison. We introduce the Mural–Illustration Similarity Prediction (MISP) model, which couples self-supervised multimodal contrastive embedding with supervised cross-modal attention fusion to produce interpretable similarity scores. MISP attains lower predictive error than single-modal and multimodal baselines and agrees closely with independent expert attributions. The murals most closely resemble the Zhou Yuejiao (周曰校) illustrated edition, identifying likely source editions and constraining when they reached Thailand. Our work extends the evidence for early Thai reception and shows how AI can advance stylistic analysis and provenance inference.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-17
DOI
https://doi.org/10.21203/rs.3.rs-11432789/v1
Primary Topic
Digital Humanities and Scholarship
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Multimodal contrastive learning traces the early dissemination of Romance of the Three Kingdoms in Thailand

李順興, Fei Yu, Hongrun Wu, 方良 et al.
Research Square
Digital Humanities and Scholarship
preprint

Multimodal contrastive learning traces the early dissemination of Romance of the Three Kingdoms in Thailand

李順興, Fei Yu, Hongrun Wu, 方良, Chou Yong Tan, Wenjie Yang, Ying Lin, Yancong Su, Baozhu Lin
preprint en

Abstract

Abstract The Romance of the Three Kingdoms (RTK, 三國演義) is a global literary canon, yet how and when it first reached Thailand remains contested because few documents survive. Using the 1838 Wat Prasert Sutthawat murals, the earliest large-scale Thai RTK visual corpus, we build the first expert-verified multimodal dataset for mural–print comparison. We introduce the Mural–Illustration Similarity Prediction (MISP) model, which couples self-supervised multimodal contrastive embedding with supervised cross-modal attention fusion to produce interpretable similarity scores. MISP attains lower predictive error than single-modal and multimodal baselines and agrees closely with independent expert attributions. The murals most closely resemble the Zhou Yuejiao (周曰校) illustrated edition, identifying likely source editions and constraining when they reached Thailand. Our work extends the evidence for early Thai reception and shows how AI can advance stylistic analysis and provenance inference.

Research Square
University of Malaya (MY), Xiamen University of Technology (CN), Minnan Normal University (CN)
National Natural Science Foundation of China, Natural Science Foundation of Fujian Province, National Office for Philosophy and Social Sciences, Division of Graduate Education
Digital Humanities and Scholarship
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon