By OpenAISourceGitHub

Manuscripts from openai/math

Fork
Clone
git clone https://openversions.com/git/index/openai-math.git
Cite
Version link
https://openversions.com/index/openai-math/commit/dfd8c46870b48511e697cbb43a3c8c91d375fa2f
Version identifier (commit hash)
dfd8c46870b48511e697cbb43a3c8c91d375fa2f

Fork to create your own copy with credit to the source. Clone to get a Git copy on your computer.

All papers

Claims

Findings

Weak pure infiniteness and O-infinity absorption

Every complex C∗C^*-algebra CC satisfying (P) (that every h∈C+h \in C_+ is properly infinite) satisfies that for all a,b∈C+a, b \in C_+, c∈Cc \in C, and ε>0\varepsilon > 0, there are s,t∈Cs, t \in C such that ∥s∗as−a∥<ε\|s^*as - a\| < \varepsilon, ∥t∗bt−b∥<ε\|t^*bt - b\| < \varepsilon, and ∥s∗ct∥<ε\|s^*ct\| < \varepsilon; consequently CC is strongly purely infinite.

If CC is an exact complex C∗C^*-algebra and there is one integer n≥1n \ge 1 such that h⊕nh^{\oplus n} is properly infinite for every h∈C+h \in C_+, then every positive element of CC is properly infinite, and consequently CC is strongly purely infinite.

If AA is a separable nuclear complex C∗C^*-algebra and there is one integer n≥1n \ge 1 such that h⊕nh^{\oplus n} is properly infinite for every h∈A+h \in A_+, then A≅A⊗min⁡O∞A \cong A \otimes_{\min} \mathcal{O}_\infty (without assumptions of unitality or simplicity).

Let CC be a nuclear complex C∗C^*-algebra. If one integer n≥1n \ge 1, fixed for the whole algebra, makes h⊕nh^{\oplus n} properly infinite for every h∈C+h \in C_+, then every positive element of CC is properly infinite.

Let CC be an exact C∗C^*-algebra satisfying (P). For every a,b∈C+a, b \in C_+ and every ε>0\varepsilon > 0, there are s,t∈Cs, t \in C such that ∥s∗a2s−a2∥<ε\|s^*a^2s - a^2\| < \varepsilon, ∥t∗b2t−b2∥<ε\|t^*b^2t - b^2\| < \varepsilon, and ∥s∗abt∥<ε\|s^*abt\| < \varepsilon; consequently CC is strongly purely infinite.

Methods

Rectangular partial isometry transport in the universal bidual C∗∗C^{**} after a spectral cutoff, preserving hereditary membership, cutoffs, Cuntz classes, and norms in every representation and at matrix levels.

Source excerpts

“Lemma 2.2 (Transport after a cutoff). Let C be a C*-algebra, let m, n ≥ 1 be integers, and let d ∈ Mn(C)+ and v ∈ Mm(C)+. Suppose d ≾ v in C ⊗ K. For ε > 0, put h = (d − ε)+. There is w ∈ Mm,n(C**) such that w*w = eh, ww* ≤ ev, (2.1) and w HerMn(C)(h) ⊂ Mm,n(C), w HerMn(C)(h)w* ⊂ HerMm(C)(v). (2.2)” · Show in paper

Graph construction of orthogonal copies in a hereditary subalgebra with fixed measurements that turn failure of nonvanishing of y(1−c∗p2c)yy(1 - c^*p^2c)y into finitely many explicit equalities.

Source excerpts

“Proposition 3.1 (Fixed measurements of a graph). Let C be a C*-algebra, let a, y ∈ C+ be contractions with a rich, and let c ∈ C˜ be a contraction. For an integer k ≥ 1 and fixed positive γg, σg satisfying (3.1), there are X ∈ Her((a − 1/8)+)+, B1, . . . , Bk ∈ Her((a − 1/8)+) with the following property.” · Show in paper

Four-entry avoidance via a directed family of finite tests and a norm ultraproduct of cyclic representations, using Blackadar–Cuntz scaling to force an isometry with a nonzero defect projection.

Source excerpts

“Proposition 4.2 (Four entries avoid the fixed measurements). Suppose C satisfies (P). Let X ∈ C+ and B1, B2, B3, B4 ∈ C. ... there exist contractions T1, . . . , T4 in Her((X − α0)+) for which no representation π of C satisfies ∥π(X)∥ > s, π(Bi) = π(Ti) (1 ≤ i ≤ 4). (4.2)” · Show in paper

Passing individual proper infiniteness (P) to bounded products to obtain a universal support contraction constant θ∈(0,1)\theta \in (0, 1), followed by iteration for arbitrarily small support compression.

Source excerpts

“Lemma 5.2 (A uniform support contraction). There exists θ ∈ (0, 1) such that for every C*-algebra C satisfying (P), every pair of positive contractions a, y ∈ C, and every contraction c ∈ C, there are positive contractions p ∈ Her(a), q ∈ Her(y) such that, for every representation π of C, ∥π(a)∥ > 1/2 =⇒ ∥π(p)∥ = 1, ∥π(y)∥ > 1/2 =⇒ ∥π(q)∥ = 1, and ∥epceq∥ ≤ θ. (5.2)” · Show in paper

Nonexact finite-tuple obstruction in a quotient of Mn(C)M_n(C) combined with Kadison transitivity and polar-decomposition merger of local rich elements.

Source excerpts

“Lemma 6.2 (A finite tuple and finite matrix approximations). There are ℓ ≥ 1, a C*-algebra W, contractions z1, . . . , zℓ ∈ W, and contractions z(r) j ∈ MNr (C), where Nr ≥ 1, r ≥ 1, and 1 ≤ j ≤ ℓ, with the following properties.” · Show in paper

Uniform perturbation on quasi-compact primitive norm slices from the universal contraction obstruction, applied in three finite coupling steps to construct square-zero elements.

Source excerpts

“Proposition 7.8 (Uniform perturbation). Let C be an exact C*-algebra satisfying (Fn) for one fixed integer n ≥ 1. ... For every T ∈ C and every η, α > 0, there is a contraction y ∈ C such that ∥(T + ηpyq)P ∥ ≥ ση for every P ∈ Prim(C) with ∥cP ∥ ≥ α. (7.14)” · Show in paper