NeurGO framework generates elite candidates for expensive optimization problems

Researchers have developed NeurGO, a novel generative framework for expensive black-box optimization problems. This approach synthesizes elite candidate solutions directly from historical population data, bypassing the need to evaluate numerous inferior options. By employing an attention-based encoder and a quality-diversity loss, NeurGO aims to capture search trends and maintain solution diversity. Experiments on standard test suites like CEC 2008 and COCO BBOB demonstrate that NeurGO achieves superior optimization performance and faster convergence within limited evaluation budgets. AI

IMPACT This framework could improve efficiency in scientific and engineering applications where computational resources are limited.

RANK_REASON The cluster contains a research paper detailing a new algorithmic framework for optimization problems. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• CEC 2008
• COCO BBOB
• MetaBBO
• Meta-Black-Box Expensive Optimization
• NeurGO

• paper
• other

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