One thing that has been worked, at least with genetic codes, is mapping them to linguistics (or rather to text), although it could be argued that the scaffoldin

One thing that has been worked, at least with genetic codes, is mapping them to linguistics (or rather to text), although it could be argued that the scaffolding of this mapping is still mathematical. I guess in this case the analogy was already there, waiting to be leveraged. Mode info: https://deepmind.com/blog/article/enformer https://deepmind.com/blog/article/enformer