CultureBot: A Framework for Retrieval-Augmented Exploration of Cultural Heritage Data
We present CULTUREBOT, an interactive framework for natural-language exploration of cultural heritage collections that combines knowledge-graph-guided retrieval, lexical-semantic retrieval, and source-linked generation.
ProofPaper ↗
Key points
- Digital cultural heritage collections encode rich relations among objects, places, periods, providers, images, and source institutions, yet users typically access them through keyword search and faceted browsing.
- CULTUREBOT maps user queries to structured catalogue relations, enriches retrieved records with textual evidence, and generates answers that link back to the underlying repository records.
- We instantiate the framework on 81,852 SearchCulture records and release an online demonstration.
- These results support graph-guided retrieval as a practical way to expose the relational structure of cultural collections through conversational access while preserving links to catalogue evidence.
Sources (1)
- [1]CultureBot: A Framework for Retrieval-Augmented Exploration of Cultural Heritage DataarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 09:00 AM
We present CULTUREBOT, an interactive framework for natural-language exploration of cultural heritage collections that combines knowledge-graph-guided retrieval, lexical-semantic retrieval, and source-linked generation.
Digital cultural heritage collections encode rich relations among objects, places, periods, providers, images, and source institutions, yet users typically access them through keyword search and faceted browsing.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 12, 2026epGen: open LoRA for LTX-2.5 that gives every sound in the video its own audio track (generation + separation, weights on HF)
- Oct 11, 2026Uncovering the Generative Landscape of Diffusion Models: A Spectral-Guided Quality-Diversity Diffusion Framework
- Oct 11, 2026AesCode: Aesthetic Code Generation with Decoupled Cross-Modal Rewards
- Oct 11, 2026Imperfect Verifiers, Exact Targets: Inference-Time Scaling for Continuous Generation
- Oct 11, 2026RSIPO: Recursive Self-Improving Policy Optimization for Agentic Reinforcement Learning via Skill-Policy Co-Evolution
- Oct 10, 2026DiffGA: Steering Frozen Diffusion Models for Materials Design via Evolutionary Search