GIST maintains evolving scholarly taxonomies by integrating LLM-extracted Related-Work partial hierarchies in a box-embedding space with novelty-aware coreset updates and budgeted hypothesized-concept retrieval, beating LLM baselines on Node/Edge Soft F1 at ~10% of their cost.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Taxonomy Maintenance In The Wild Over Evolving Scholarly Data: Reliability, Efficiency, and Cost-Effectiveness
GIST maintains evolving scholarly taxonomies by integrating LLM-extracted Related-Work partial hierarchies in a box-embedding space with novelty-aware coreset updates and budgeted hypothesized-concept retrieval, beating LLM baselines on Node/Edge Soft F1 at ~10% of their cost.