Q18 — Can spin structure explain disentangler bounds?

[marwahaha]

Discussion of Q18: Can spin structure explain disentangler bounds?.

For approximate disentanglers, relate input dimension to the requirements that outputs be nearly separable and that every separable target be approximately attainable.

Possible reductions, counterexamples, special cases, and related results are welcome. Please state the assumptions and link to relevant work.

The question page contains the full formulation and references. This conversation also appears beneath the question on both versions of the website.