Researchers have developed MACE, a novel Multi-Agent Candidate Event acquisition method designed to improve event linking in text. This approach refines event structure before linking by employing specialized LLM agents to gather evidence for time, location, participants, and event types. MACE then exposes intermediate queries to candidate-event lookup tools and allows a coordinator to revise the evidence set before constructing final candidates. Experiments on two benchmarks demonstrated that integrating MACE consistently enhances accuracy across various event linking models, proving its effectiveness in improving candidate acquisition without altering the core linking model. AI
IMPACT Improves accuracy in event linking tasks by refining candidate acquisition through multi-agent LLM collaboration.
RANK_REASON The cluster contains a research paper detailing a new method for AI event linking. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
