A multi-agent LLM system using iterative planning and combinatorial innovation produces NLP research ideas with greater diversity and novelty than prior baselines, with quality between accepted and rejected conference papers.
For evaluation, we combined cross-group Swiss-system pairing with zero-shot LLM-based pairwise com-parison
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Enhancing Research Idea Generation through Combinatorial Innovation and Multi-Agent Iterative Search Strategies
A multi-agent LLM system using iterative planning and combinatorial innovation produces NLP research ideas with greater diversity and novelty than prior baselines, with quality between accepted and rejected conference papers.