AFTD Auto-Formalizing Theoretical Domains

mutual_power_divergence_subadditive_imp_golden_ratio_le

theoremverifiedoriginal

For any positive exponent beta, if the mutual power-divergence information with generator f_beta is subadditive under all finite Markov chains Y <- X -> Z, then beta >= (Real.sqrt 5 - 1) / 2.

Statement

theorem mutual_power_divergence_subadditive_imp_golden_ratio_le (β : ℝ)
    (hβ0 : 0 < β)
    (hsub : mutual_f_information_subadditive (power_divergence_generator β)) :
    (Real.sqrt 5 - 1) / 2 ≤ β
Related work
arXiv:2610.21231 'On the Subadditivity of Mutual f-Information'
Verified
31 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 3
Used by 1

Read back from the Lean

The theorem asserts that for any real number β > 0, if the f-divergence generator f(x) = (x^β - β*x + (β - 1)) / (β * (β - 1)) makes mutual f-information subadditive for all conditionally independent channels over finite types, then β is at least (√5 - 1) / 2 (the reciprocal of the golden ratio).

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

The proof, explained

The proof specializes the general subadditivity hypothesis to the case where the alphabets are binary, X = Y = Z = Fin 2. It then invokes the binary version of the theorem, `mutual_power_divergence_binary_subadditive_imp_golden_ratio_le`, which already establishes that subadditivity on binary alphabets forces the exponent β to be at least the golden ratio conjugate (√5 - 1) / 2.

  1. (√5 - 1) / 2 ≤ β

    The hypothesis `hsub` guarantees subadditivity of the mutual power-divergence information across all finite alphabets. Specializing `hsub` to binary alphabets (setting X = Y = Z = Fin 2) gives the binary subadditivity condition. Applying the theorem `mutual_power_divergence_binary_subadditive_imp_golden_ratio_le` to β, the positivity assumption hβ0, and this binary specialization immediately yields the bound (√5 - 1) / 2 ≤ β.

    line 5

    uses mutual_power_divergence_binary_subadditive_imp_golden_ratio_le

        (Real.sqrt 5 - 1) / 2 ≤ β := mutual_power_divergence_binary_subadditive_imp_golden_ratio_le β hβ0 (hsub (Fin 2) (Fin 2) (Fin 2))
mutual_power_divergence_binary_subadditive_imp_golden_ratio_le
If β > 0 and the mutual power-divergence information is subadditive for binary alphabets (Fin 2), then β ≥ (√5 - 1) / 2.

Written by a model (gemini-3.8-flash) from the Lean source alone (no goal states were available). Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

/-- Subadditivity of mutual power-divergence information across all finite alphabets implies beta >= (sqrt 5 - 1) / 2. -/theorem mutual_power_divergence_subadditive_imp_golden_ratio_le (β : ℝ)    (hβ0 : 0 < β)    (hsub : mutual_f_information_subadditive (power_divergence_generator β)) :    (Real.sqrt 5 - 1) / 2 ≤ β := mutual_power_divergence_binary_subadditive_imp_golden_ratio_le β hβ0 (hsub (Fin 2) (Fin 2) (Fin 2))
See it in the graphDiscuss this resultChallenge it