Introduce reproducibility challenge, how it is underestimated in the community, there is barely no framework available. Talk about the stochasticity of the proposed approach and how it could be a potential problem for reproducibility.
Present the organization of the paper (different sections
Check the papers that Dave linked in another issue