MOSAIC: Adversarial Co-evolution of Specialist Heuristics and Problem Instances for LLM-based Automated Heuristic Design
The episode discusses MOSAIC, a paper by Gungordu et al. that uses LLMs to design heuristics for automated problem-solving. The system employs adversarial co-evolution to create a portfolio of specialist heuristics and problem instances, moving beyond single algorithms by generating rich insights and diverse test cases.