Task-free Adaptive Meta Black-box Optimization

ICLROral2026

Authors
Chao Wang, Licheng Jiao, Lingling Li, Jiaxuan Zhao, Guanchun Wang, Fang Liu, Shuyuan Yang
Affiliation
Xidian University
Venue
ICLR 2026
Track
Oral

TL;DR

Handcrafted optimizers become prohibitively inefficient for complex black-box optimization (BBO) tasks. MetaBBO addresses this challenge by meta-learning to automatically configure optimizers for low-level BBO tasks, thereby eliminating heuristic dependencies.

Opening excerpt from the authors’ abstract. source

Read the paper

Topics

meta-learning optimization efficient

← All ICLR 2026 Oral papers · Browse the whole archive