The exact capacity under a spike budget is the lower binomial sum.
ProvedNeuralCoding.SparseEnergyTradeoff.card_budgetCodebookaether-catalognovelty
The exact capacity under a spike budget is the lower binomial sum.
theorem NeuralCoding.SparseEnergyTradeoff.card_budgetCodebook(N k : ℕ) :
(budgetCodebook N k).card = ∑ j ∈ Finset.range (k + 1), N.choose j := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Novelty/SparseEnergyTradeoff.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Novelty/SparseEnergyTradeoff.lean import Mathlib import Definitions.Def_Novelty_NeuralCoding import Definitions.Def_Novelty_SparseEnergyTradeoff /-! # Sparse neural coding: exact capacity and an energy–information bound A binary pattern on `N` neurons has energy equal to its Hamming weight. This file studies the codebook consisting of all patterns using at most `k` spikes. Its size is exactly the Hamming-ball volume `∑ j ∈ range (k+1), N.choose j`. For exactly `k` spikes, the number of concepts is `N.choose k`. The central energy theorem proves that this capacity is at most `N^k`; consequently its information content per spike is at most `log₂ N`. Thus fixed sparsity gives polynomial raw capacity while information per unit energy grows only logarithmically with population size. The result also applies directly to a one-percent budget by substituting `k = N / 100`. -/ open NeuralCoding.SparseEnergyTradeoff open Finset open NeuralCoding
Formal statement
theorem NeuralCoding.SparseEnergyTradeoff.card_budgetCodebook(N k : ℕ) :
(budgetCodebook N k).card = ∑ j ∈ Finset.range (k + 1), N.choose j := by sorrySource