Robustness and Generalization in Algorithm Discovery

Methods that reduce overfitting through diverse benchmarks, adversarial instances, validation protocols, co-evolution, and out-of-distribution testing