LLMs extract user preferences for knowledge graphs and recommendations

Researchers have developed a method to extract structured user preference triples from conversational data using Large Language Models, aiming to build Personal Knowledge Graphs (PKGs). This approach converts unstructured "strings" into semantic "things" by linking extracted triples to Wikidata identifiers. Evaluations using Qwen- and Gemma-based models showed that some models performed well in semantic extraction and demonstrated utility in downstream recommendation tasks. AI

IMPACT This research could improve personalized recommendation systems by enabling more accurate extraction of user preferences from conversational data.

RANK_REASON The cluster contains an academic paper detailing a new method for knowledge graph construction using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Gemma
• Large Language Models
• Personal Knowledge Graphs
• Qwen
• RDF
• Wikidata

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →