Ah! That clears it up. You then mean Deep Learning has a lot in common with differential geometry and manifolds in general. That I will definitely agree with.

Ah! That clears it up. You then mean Deep Learning has a lot in common with differential geometry and manifolds in general. That I will definitely agree with. DG and manifolds have far richer and informative structure than topology.