Can purification give a representation-theoretic adversary method?
Purify an oracle ensemble into an inaccessible register, decompose it into representation sectors, and bound the change of a progress observable per query.
Discuss this questionChoose a group-symmetric oracle ensemble and a decision or state-generation task. Purify the hidden instance in an inaccessible register, use representation theory to organize that register, and construct a progress observable whose initial value, required successful final value, and maximum change per oracle query yield a lower bound. Can this recover or strengthen a known bound in a form native to subsystem spin?
The method class already exists: symmetry-assisted adversaries and compressed/recording oracles are established. The proposed contribution is a subsystem-spin formulation or new application. Sector weights alone may discard multiplicity and coherence information needed for a sharp bound. Purification is an analysis device, not a register the algorithm gets to manipulate.
Why this is interesting
Purification, compressed oracles, and symmetry-assisted adversaries provide established starting points. A subsystem-spin interpretation could make their common structure easier to use.
A place to start
Choose one small - or permutation-covariant discrimination task; compute the hidden-register sector weights after one and two queries, then compare candidate progress bounds.
Related work
These works provide context and established results. They need not be the origin of this exact formulation.
Discuss this question
Reading is public. Posting requires a GitHub account. You can also write to Kunal.