Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
In: Proceedings of IEEE International Conference on Evolutionary Computation
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
method 1
citation-polarity summary
roles
method 1polarities
use method 1representative citing papers
DRSR uses Quality-Diversity to produce diverse symbolic regression expressions differing in residual distributions, enabling post-search selection on synthetic and astronomical data.
Drift analysis on a mixed-integer benchmark shows (1+1)-LB-ES risks premature convergence with large numbers of integer variables while (1+1)-LUB-ES achieves linear convergence after integers are fixed under suitable bounds.
Safe ASNG uses Walsh-function surrogate models to estimate Lipschitz constants of safety functions and projects new binary solutions onto safe regions around previously evaluated safe points to suppress unsafe evaluations during optimization.
An empirical study of JEPA world models identifies architecture, training objective, and planning choices that yield a model outperforming DINO-WM and V-JEPA-2-AC on navigation and manipulation tasks.
WA-ASNG augments ASNG with weight adaptation that maximizes an estimated update signal, showing improved results over PBIL and ASNG on binary problems with population sizes 25-100 and under noise.
Shallow MLPs and dense CPGs outperform deeper MLPs and Actor-Critic RL in bounded robot control tasks with limited proprioception, with a Parameter Impact metric indicating extra RL parameters yield no performance gain over evolutionary strategies.
citing papers explorer
-
Benefits of Low-Cost Bio-Inspiration in the Age of Overparametrization
Shallow MLPs and dense CPGs outperform deeper MLPs and Actor-Critic RL in bounded robot control tasks with limited proprioception, with a Parameter Impact metric indicating extra RL parameters yield no performance gain over evolutionary strategies.