Statement
Problem 4.3 of Goto et al.: an “inverse Hayden–Preskill” question — rather than recovering an input state from radiation given a known interior unitary, can features of the interior dynamics themselves be inferred from the emitted radiation?
Why It Matters
Peripheral to the group; a learning-theory / black-hole question. Included for completeness of Theme 3.
What We Know
- Standard Hayden–Preskill assumes the interior unitary is known.
- Hamiltonian learning and process tomography give finite-dim analogs of “learning the dynamics from outputs.”
Possible Approaches
- Frame as quantum process learning with restricted (radiation-only) access.
- Bound learnability by scrambling / design properties (links to many-body-dynamics-for-near-haar-encoding).