Evolution meta-data

[csvelayutham]

I am currently using the EvoGym datasets (both the GitHub release and the Hugging Face version) for research on robot morphology optimization. I have a question regarding the structure of the dataset.

At the moment, each robot entry includes fields such as uid, env_name, generated_by, body, connections, and reward. However, there is no information indicating: the independent evolutionary run a robot was generated in, the run seed, the generation number, the population index, or any lineage or evolutionary trace.

For research involving comparisons across runs, trajectory visualization, and analysis of optimization dynamics, run-level metadata is essential. Without run identifiers, it is not possible to reconstruct independent trials or evaluate per-run progression.

Could you please clarify:

Were the original results produced using multiple independent evolutionary runs per environment and per algorithm?

If so, is the run ID / seed / generation trace available internally?

Would it be possible to release this information, or alternatively, share a version of the dataset annotated with run identifiers?

If such data cannot be released, is there any recommended way to recover run structure from the existing dataset?

Having access to run-level metadata would significantly improve the scientific reproducibility and analysis potential of the EvoGym dataset.

Thanks