Mechanisms by which biological evolution generates novelty

Investigate and characterize the mechanisms by which biological evolution generates novelty in traits and structures, beyond current computational approaches such as novelty search and quality-diversity algorithms.

Background

The paper situates its contribution within broader questions about open-endedness and novelty in evolutionary systems. The authors explicitly state that fundamental questions persist about how biological evolution produces novelty, referencing existing computational methods that address parts of the problem while acknowledging remaining gaps.

This establishes a high-level open question that motivates the development of diffusion-model-based evolutionary strategies and relates to the broader theoretical understanding of generative mechanisms in biological evolution.