Jcost_pos_of_ne_one
plain-language theorem explainer
The lemma proves that the J-cost function is strictly positive for every positive real not equal to one. Researchers modeling speech intelligibility, chemical thresholds, and climate predictability cite it to guarantee positive penalties away from the symmetric scale. The proof rewrites the cost via its squared algebraic form and applies standard positivity of squares and positive denominators.
Claim. For every real number $x > 0$ with $x ≠ 1$, the J-cost satisfies $J(x) > 0$, where $J(x) = (x + x^{-1})/2 - 1$.
background
The J-cost is defined by Jcost(x) := (x + x^{-1})/2 - 1. This is the unique function satisfying the Recognition Composition Law that vanishes at the fixed point x = 1. An equivalent form J(x) = (x-1)^2/(2x) for x ≠ 0 is obtained by field simplification and ring algebra, as recorded in the sibling lemma Jcost_eq_sq.
proof idea
The proof first obtains x ≠ 0 from the positivity hypothesis. It rewrites via the upstream lemma Jcost_eq_sq, then applies div_pos. The numerator is positive by sub_ne_zero followed by sq_pos_of_ne_zero; the denominator is positive by mul_pos of 2 and x.
why it matters
This lemma is invoked in more than forty downstream results, including srCost_pos_off_threshold for speech, aestheticCost_pos_off_optimum for cultural aesthetics, and oxidative_stress for reactive oxygen species. It supplies the strict positivity half of J-uniqueness (T5) and thereby supports the forcing chain from the Recognition Composition Law through the eight-tick octave to D = 3. It closes the zero-at-unity property required by the phi-ladder mass formula.
Switch to Lean above to see the machine-checked source, dependencies, and usage graph.
papers checked against this theorem (showing 30 of 45)
-
Variational PINN models fracture cracks with higher accuracy
"non-convex phase-field landscape may converge to incorrect crack paths"
-
Kuramoto flow fully described by Morse theory
"V(Θ)=(1−|Z|^2)m²/2 … minimum of V is zero … maximum m²/2"
-
New decoder lifts VLA robot success from 40.4% to 50.2%
"stationarity-executability gap: the iterate with the smallest residual need not execute best"
-
Multichannel divergence loss trains DNNs for beamforming
"L2 = ∑ tr(Xt,f X̂t,f⁻¹) + log det(X̂t,f)"
-
1% strain doubles ion conductance through graphene pores
"colossal – 100 % – change in conductance ... small pore strains – 1 % ... optimal transport in a primarily diffusive regime"
-
EDM data bounds Lorentz-violating coefficients at 10^{-31} (eV)^{-2}
"Electric dipole moments of atoms can arise from P-odd and T-odd electron–nucleon couplings."
-
Axioms single out Gini and Atkinson indices for egalitarian ethics
"ratio-invariable is generally regarded one of the strongest and reasonable condition for distributive theory"
-
LLMs default to positive average personas in behavior simulations
"Real human behavior is inherently sparse, with positive interaction rates remaining below 10%. By contrast, all evaluated LLM-based simulators exhibit a hyper-activity bias."
-
Crystal polaritons raise quantum-light efficiency by orders of magnitude
"J_{h,h'}(k∥,ω) = μ₀ω²/πℏ … n(h)·G_AH(k∥,r₀^{(h)},r₀^{(h')},ω)·n^{(h')}"
-
Self-limiting losses compress embeddings without overfitting
"a self-limiting hinge-based triplet loss produces zero gradient once a triplet satisfies the margin constraint, bounding the total perturbation applied to the pretrained embedding space... the fraction of active triplets drops below 10% within 5–15 epochs"
-
Lyapunov descent in doubling crystal stabilizes laser intensity
"The intensity-channel mechanism is intuitive. The crystal depletes intensity as I², so brighter time-slices lose proportionally more energy than dim time-slices, flattening I(t) toward its mean."
-
Training augmentations alone match FGIR accuracy without crops
"data-aware augmentations during training only can enable a model to achieve excellent accuracy even without crops, significantly reducing inference costs"
-
Normal-aligned Jacobians explain success on sparse data
"normal-aligned classifiers ... implement the program of an optimal match filter on the training data"
-
Smart charging cuts EV grid upgrade costs on real feeders
"LB outperforms TOU when the customer enrollment levels are high"
-
Progress rewards let agents escape noisy-TV traps
"r_i_t = E_D[ε^{(τ-1)}_t(o_{t+1})] - ε^{(τ)}_t(o_{t+1})"
-
Local positivity yields turnpike in mean field games
"−˙Φ(t)≥σ|Φ(t)|... exponential decay... turnpike property"
-
Sentipolis adds lasting emotions to LLM social agents
"s(t+Δt)=s(t) 2^(-Δt/T½) half-life decay"
-
Hořava-Witten on joined circles dual to 0B orientifold SO(16)^4
"Hořava-Witten theory on S¹∨S¹ … SO(16)⁴ … Fabinger-Hořava configurations … bifundamental tachyons/fermions at junction"
-
Dissipation potentials enable PDE-free tracking in port-Hamiltonian systems
"Condition (13) is the scalar analogue of positive definiteness of the damping matrix... When Fp(p)=½(∇pHqp)⊤Rp∇pHqp ... system (11) reduces to the standard PHS"
-
LLM faults expose 99% worse RMSE in edge lane following
"ResNet18 lane-following model... R^2 ≈0.85 on clean data... RMSE increasing by up to 99%... within-0.10 localization accuracy dropping to as low as 31.0% under fog"
-
LLMs hold memories in entangled states needing active suppression
"removing letter-identity information modestly improves N-back performance... providing causal support for representational interference"
-
Dynamics emerge from relaxing local gauge mismatches
"rij ≡ ϕ (Tij Uj − Ui) ... L[U, W] = ½ Σ ⟨Ri, Ri⟩ + λ/2 Σ ⟨Wij, Wij⟩"
-
Additive composition structures latent actions for robot learning
"suppress information that does not compose additively: static environment terms and future leakage"
-
RL reasoning performance saturates as policy entropy drops to zero
"the change in policy entropy is driven by the covariance between action probability and the change in logits, which is proportional to its advantage"
-
Non-CARA preferences leave prices only partially revealing
"the probability average of a logistic transform is not the logistic transform of the average ... Jensen gap ... is the size of the revelation failure"
-
Spiking transformer filters predictable signals for 80.5% ImageNet accuracy
"SAFformer actively suppresses predictable signals and focuses on salient visual features... higher-level cortical areas form a holistic prediction... only the significant discrepancies are propagated upward as feedback signals"
-
Caution reduces reward hacking in Best-of-N sampling
"linear setting... α(y) approximates projection onto V^⊥... E[r⋆(yipess)−r⋆(yˆi)] ≳ √log N"
-
Quality-Aware Exploration Budget Allocation for Cooperative Multi-Agent Reinforcement Learning
"Successor Distance d^π_SD(x,y) = log(p^π_γ(s_f=y|s_0=y) / p^π_γ(s_f=y|s_0=x)) — a learned quasimetric."
-
Gradient alignment preserves client directions in federated learning
"γt ← 0.5 − 0.5 erf(|mt|/√(2(vt−mt²)+ε)) sign(−mt·Δrk)"
-
XXZ spin conductivity scales as 1/T at high temperature
"We show that the spin conductivity is proportional to 1/T in the limit T→∞ and its high temperature proportionality constant diverges in the case where one-quasiparticle magnetization is infinitely large."