Skip to content

PEPS FT: formalize the generalized two-injective-tensor comparison #1361

Description

@LionSR

Goal

Formalize the generalized two-injective-tensor comparison used in arXiv:1804.04964, Section 3, Lemma inj_equal_tensors_2.

The paper proves that if two pairs of injective tensors agree after inserting an arbitrary matrix on every shared virtual bond, then the corresponding tensors are proportional with inverse scalar factors:

A_1 = lambda * B_1,
A_2 = lambda^{-1} * B_2.

The scalar-reduction substep in this proof is tracked separately by #1362. This issue is the enclosing two-tensor comparison used after eq:inj_equal_edge from #1364, before the final one-vertex-versus-complement specialization tracked by #1360.

Blueprint target

The current worktree adds a not-yet-formalized blueprint node:

thm:peps_twoInjectiveTensorInsertionComparison

and a semantic diagram macro:

\TNPEPSTwoInjectiveTensorInsertionComparison

The diagram depicts two injective tensors joined by several shared virtual bonds, with an arbitrary matrix X inserted on one such bond, matching the displayed hypothesis of lem:inj_equal_tensors_2.

The proof substep #1362 is represented by:

thm:peps_twoInjectiveGaugeScalarReduction
\TNPEPSTwoInjectiveGaugeScalarReduction

Dependencies

Acceptance criteria

  • State the Lean theorem corresponding to thm:peps_twoInjectiveTensorInsertionComparison with arbitrary matrix insertions on shared virtual bonds.
  • The theorem assumes injectivity of both tensor blocks, not merely equality of fully contracted scalar states.
  • The conclusion gives reciprocal scalar proportionality of the two tensors.
  • The proof follows the argument of lem:inj_equal_tensors_2, using the scalar-reduction substep PEPS FT: formalize scalar reduction in the two-injective-tensor comparison #1362 to reduce the possible remaining virtual gauges to scalar multiples of the identity.
  • No new untracked sorry is introduced outside this theorem endpoint.
  • Chapter 13a keeps the diagram attached to this theorem node.

Metadata

Metadata

Assignees

Labels

1804.04964arXiv:1804.04964 - FT for normal tensor networksfollow-upformalizationLean 4 formalization task

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions