JCDL 2026 · Tutorial
AI for Organizing Digital Collections
Practical approaches with large language models and vision-language models for classification, subject analysis, and archival image description.
Date
October 16, 2026
Format
Hybrid · 3 hours
Location
Dallas, Texas + online
About the tutorial
Turn emerging AI capabilities into responsible collection workflows.
Digital libraries, museums, and cultural heritage institutions manage growing collections of text, images, and multimodal materials. Organizing these resources remains labor-intensive, while metadata backlogs and inconsistent description can limit discovery.
This tutorial connects core model concepts with practical applications. Through accessible explanations, demonstrations, and guided exercises, participants will use LLMs for textual metadata tasks and VLMs for archival image description—then examine how to evaluate, refine, and responsibly integrate their outputs.
Three-hour program
Schedule
| Time | Section |
|---|---|
| 0–10 min | Welcome & introduction: background, motivation, and AI impact |
| 10–45 min | Part 1: LLMs for classification, RAG, and prompt engineering |
| 45–60 min | Environment setup for hands-on practice |
| 60–90 min | Hands-on 1: applying LLMs to textual metadata |
| 90–105 min | Coffee break |
| 105–135 min | Part 2: VLM capabilities, image description, and semantic analysis |
| 135–165 min | Hands-on 2: applying VLMs to archival images |
| 165–180 min | Ethical considerations and workflow integration |
Open resources
Tutorial materials
Download the slides, launch the hands-on notebooks, and explore the accompanying code and readings.
Colab notebooks
Dataset & code
Reading list
Suggested reading
Related papers
-
A Hybrid Framework for Subject Analysis: Integrating Embedding-Based Regression Models with Large Language Models
Liu, Song, Zhang, Thomale, He, and Hong · 2025
-
An Experiment with the Use of ChatGPT for LCSH Subject Assignment on Electronic Theses and Dissertations
Chow, Kao, and Li · 2024
-
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Lewis et al. · NeurIPS 2020
-
QLoRA: Efficient Finetuning of Quantized LLMs
Dettmers, Pagnoni, Holtzman, and Zettlemoyer · NeurIPS 2023
-
LoRA: Low-Rank Adaptation of Large Language Models
Hu et al. · ICLR 2022
-
Learning Transferable Visual Models From Natural Language Supervision
Radford et al. · ICML 2021
-
Multimodal Metadata Assignment for Cultural Heritage Artifacts
Rei et al. · Multimedia Systems, 2023
-
Responsible AI Practice in Libraries and Archives: A Review of the Literature
Mannheimer et al. · Information Technology and Libraries, 2024
This work is supported by the Institute of Museum and Library Services (IMLS) under Grant No. LG-256666-OLS-24. Additional support was provided through IMLS Grants LG-256661-OLS-24 and RE-256673-OLS-24.