Student Forum
The programme is still being finalized and is subject to ongoing updates as sessions are scheduled. Please check back regularly for the latest changes.
Authority Data as Context-Bearing Entities in Bibliographic Ontology Environments
Authority data is traditionally treated as a tool for identifying and describing entities with precision—a reliable but largely technical access point. This presentation reconsiders authority data as an autonomous, context-bearing entity whose significance extends beyond its descriptive function. It shows that authority data carries context across four dimensions—relational, temporal, epistemic, and domain-specific—that together form a distinct mode of contextual meaning. Here, contextualization appears not as incidental but as an intrinsic property of authority data itself. On this basis, the presentation suggests a shift away from treating authority data as a mere access point and toward understanding it as a context-bearing entity embedded within a structured knowledge network.
-
Seunghui Lee
PhD Student
Chung-Ang University
Seunghui Lee is a PhD student in Knowledge Organization at Chung-Ang University in Seoul, South Korea. He has worked as a librarian at the National Library of Korea. His research interests include knowledge organization, metadata, and bibliographic ontology. He focuses on how knowledge organization systems (KOSs) reflect the ways in which society organizes and shares knowledge, seeking to bridge theoretical inquiry and practical implementation.
From Genre Categorization to User Discovery: Knowledge Organization of Online Comic Platforms in Taiwan
Online comic readers rely on a wide range of descriptions to find new works, yet these user-centric terms are often different from the categorization terms used by platforms. Moreover, due to the subjective nature of “aboutness” judgments, the same comic work is often categorized under different themes on different platforms, creating a gap between the platform’s categorization and users’ natural language descriptions. This master thesis study aims to explore the current status of genre categorization on online comic platforms in Taiwan and compare them with readers’ perspectives. After collecting terms from popular Taiwanese platforms and performing content analysis, interviews will be conducted to gain a deeper understanding of user feedback. Preliminary analysis shows that platform category terms are multifaceted, encompassing mood, setting, and audience. Additionally, semantic inconsistencies between platforms, and between different language versions of the same platform, are also observed. Furthermore, a platform’s background (e.g., quasi-governmental vs. commercial) leads to significant differences in organizing logic, increasing the cognitive load for readers. Through this DCMI Student Forum, the author expects to discuss and receive feedback on how to integrate readers’ perspectives into existing comic genre categorization.
-
Tzu-Yun Chien
National Taiwan University
Tzu-Yun Chien is a Master’s student in the Department of Library and Information Science at National Taiwan University. Her research interest focusing on human–computer interaction and information behavior. Her prior research on user behaviors in AI-assisted tasks has been published as a full conference paper. Her ongoing master's thesis focuses on the differences between existing genre categorization frameworks, platform labeling practices, and user interpretations within online comics.
Korean E-Book Metadata Enhancement & Withdrawal Project: Promoting Digital Collections & Environmental Sustainability
Authors: Donghyun Kim
The Korean E-Book Metadata Enhancement & Withdrawal Project at the Hamilton Library in the University of Hawaiʻi at Mānoa addressed two interconnected challenges: limited shelf space in the Asia Collection and insufficient discoverability of newly acquired Korean e-books. Conducted from June to August 2025, this pilot project focused on enhancing e-book metadata in Alma using OCLC WorldCat records and identifying corresponding low-circulation physical volumes for withdrawal. The project enhanced 350 e-book records and withdrew 257 print volumes based on loan and in-house use data. The workflow involved Korean-script title searching, MARC record overlay, editing of key bibliographic fields, and careful review of withdrawal candidates. Challenges included inconsistent romanization, limited subject and genre metadata, and the difficult decisions about balancing preservation and space management. The project demonstrates how targeted metadata enhancement can improve digital resource discoverability while supporting environmentally sustainable library practices. It also highlights the importance of linguistic expertise, inclusive metadata practices, and future batch cataloging workflows for continuing Korean e-book access initiatives.
-
Donghyun Kim
LIS Student
University of Hawaiʻi at Mānoa
Donghyun Kim is a master’s student in the Library and Information Science program at the University of Hawaiʻi at Mānoa. Before joining the program, he worked as a theater director for seven years in South Korea. He currently works at the Center for Korean Studies as a Graduate Assistant, where he preserves archival materials and provides technical assistance in appropriate cultural protocols and professional expertise. He pursues academic librarianship specializing in cataloging and metadata, who advocates for access and equity with inclusive library technologies.
Modeling Bibliographic Uncertainty in Korean Rare Books Using RDF-Star
Korean rare books pose a persistent challenge for bibliographic description: authorship, publication date, and custodial history are frequently uncertain or recoverable only through indirect inference. Existing catalog practice handles this through placeholder notation or omission, discarding the contextual information—source type, confidence level, inferential basis—that makes an assertion interpretable. This paper proposes RDF-Star, the W3C extension now entering the RDF 1.2 specification (W3C, 2024), as a practical mechanism for representing structured bibliographic uncertainty at the triple level. We demonstrate the approach through two cases drawn from Korean historical collections: the undated Jukchangjip manuscript, whose compilation date is inferred through textual comparison with a later edition, and the five-century custodial history of Weolincheongangjigok, where each transfer event differs in documentary support and reliability. Both cases are validated in a working implementation using pyoxigraph 0.5.8, confirming that a compact controlled vocabulary of source types and confidence levels is sufficient to support structured retrieval of uncertain metadata. The approach layers onto existing catalog data without replacement, making incremental adoption feasible for institutions with large legacy collections.
-
Inyoung Jung
Kyungpook National University
I am a student at Kyungpook National University in South Korea, focusing on metadata for Korean rare books and cultural heritage materials. My previous work includes studies on archival metadata, cataloging of clan-held collections, and comparative analysis of rare book cataloging rules. I am currently researching practical cataloging workflows for Korean rare books, with a strong interest in how metadata standards and linked data technologies can better represent the complexity and uncertainty inherent in historical bibliographic records.
Multi-Dimensional Evaluation of Semantic Retrieval Adequacy in a Neuro-Symbolic Heritage Knowledge Graph
Authors: Nuria Ferran-Ferrer, Miquel Centelles, Matheus Jenevain
Retrieval over heritage knowledge graphs is relational and ontology-dependent: queries can execute yet deliver weak interpretation, especially where archival silences recur as searchable absences. This process report presents a procedural framework for evaluating semantic retrieval adequacy in HerStory NeSy AI— repression–era neuro-symbolic platform with Wikidata-style `Q`/`P` commitments in Neo4j Aura. A six-dimensional protocol spans schema quality, cross-channel robustness (UI, API, Cypher, SPARQL), fairness readiness, interpretability, governance, and `GroundTruthMapping` alignment. Each intent runs thrice per snapshot (two manual platform sessions; one scripted pass pairing browser MCP with Aura read-Cypher), enabling contrast between interaction-layer behaviour and graph-grounded reads on the same analytic intent. First version runs (2,890 nodes; 4,482 relationships) show SPARQL/API mediation failures, 51.04% `P`-code penetration with 0% authorised `P21` on persons, thin UI payloads despite in-graph corpora, and blocked provenance probes—demonstrating why retrieval and mapping must be evaluated jointly rather than as isolated accuracy checks.
-
Elena María Gómez Rico
Predoctoral student
Universitat de Barcelona
Elena M. Gómez-Rico is a predoctoral researcher in Information and Communication at the Universitat de Barcelona (HerStory NeSy AI, PID2023-147673OB-I00). She holds degrees in Product Design Engineering and Humanistic Studies (EAFIT, Colombia) and an Official Master’s in Digital Humanities (UB). Her thesis develops participatory, user-centred information architectures and neurosymbolic AI for digital historical archives, addressing epistemic justice and gendered representation in GLAM. She has published on design pedagogy, machine-mediated content and public science communication.
Hosted by
