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

The Kirchberg–Rørdam character criterion

For any nonzero unital separable complex C∗C^*-algebra AA and any free ultrafilter ω\omega on N\mathbb{N}, the norm central-sequence algebra Fω(A)F_\omega(A) has no characters if and only if A≅A⊗min⁡ZA \cong A \otimes_{\min} \mathcal{Z}.

For every nonzero unital complex C∗C^*-algebra DD without characters, there is an integer m≥1m \geq 1 and a unital ∗*-homomorphism I(2,3)⟶D⊗max⁡mI(2, 3) \longrightarrow D^{\otimes_{\max} m}, where I(2,3)={f∈C([0,1],M2⊗M3):f(0)∈M2⊗13,f(1)∈12⊗M3}I(2, 3) = \{f \in C([0, 1], M_2 \otimes M_3) : f(0) \in M_2 \otimes 1_3, f(1) \in 1_2 \otimes M_3\}.

For any nonzero unital separable complex C∗C^*-algebra DD without characters, its infinite minimal tensor power B=D⊗min⁡∞:=lim→⁡n≥1(D⊗min⁡n,x↦x⊗1D)B = D^{\otimes_{\min} \infty} := \varinjlim_{n \geq 1} (D^{\otimes_{\min} n}, x \mapsto x \otimes 1_D) satisfies B≅B⊗min⁡ZB \cong B \otimes_{\min} \mathcal{Z} and contains a unital copy of Z\mathcal{Z}.

For every nonzero unital C∗C^*-algebra DD without characters, there are an integer m≥1m \geq 1, the maximal tensor power P=D⊗mP = D^{\otimes m}, and a unital ∗*-homomorphism η:C→P\eta : C \to P such that η(ιe11)\eta(\iota e_{11}) is full in PP, where C={f∈C([0,1],M2):f(0)∈C1M2}C = \{f \in C([0, 1], M_2) : f(0) \in \mathbb{C}1_{M_2}\} and ι(t)=t\iota(t) = t.

If DD is a nonzero unital C∗C^*-algebra without characters, then some finite maximal tensor power PP of DD contains a full element ww satisfying w2=0w^2 = 0 and ∥w∥=1\|w\| = 1.

Let PP be a nonzero unital C∗C^*-algebra. If there is a unital ∗*-homomorphism η:C→P\eta : C \to P such that η(ιe11)\eta(\iota e_{11}) is full in PP, then for some finite m≥1m \geq 1, B=P⊗mB = P^{\otimes m} contains orthogonal positive contractions b1,b2b_1, b_2 and vB∈Bv_B \in B with b1=vB∗vBb_1 = v_B^* v_B, b2=vBvB∗b_2 = v_B v_B^*, and 1B−b1−b2≾(b1−1/2)+1_B - b_1 - b_2 \precsim (b_1 - 1/2)_+, and consequently there is a unital ∗*-homomorphism I(2,3)⟶P⊗mI(2, 3) \longrightarrow P^{\otimes m}.

Methods

Representation-stable tensor labels are constructed via Sard's theorem on P×Sr−1×Wr→(R2)⊗Rn\mathcal{P} \times S^{r-1} \times W^r \to (\mathbb{R}^2)^{\otimes_{\mathbb{R}} n} so that they remain linearly independent under arbitrary rank-at-least-two maps on individual tensor legs and under multiplication of commuting von Neumann factors.

Source excerpts

“Lemma 4.4. Let V be a finite-dimensional real vector space with m = dimR V ≥ 2. For every integer r ≥ 1, there are n ≥ 1 and tensors x1, . . . , xr ∈ V⊗Rn with the following property: whenever Pj : V → Ej are real linear maps of rank at least two, the tensors (P1 ⊗ · · · ⊗ Pn)(xi), 1 ≤ i ≤ r, are linearly independent over R.” · Show in paper

“Lemma 4.1. Let M ⊆ B(H) be a von Neumann factor with 1M = 1H, and let B ⊆ M′ be a unital algebra. The multiplication map M ⊙ B −→ B(H), a ⊗ b |−→ ab, is injective.” · Show in paper

A unital matrix-cone homomorphism η:C→P\eta : C \to P with full first corner is constructed from a full square-zero element w∈Pw \in P via polar decomposition w=u∣w∣w = u|w| in P∗∗P^{**} and commuting functional calculus with h=∣w∣+∣w∗∣h = |w| + |w^*|.

Source excerpts

“Commuting functional calculus now defines a ∗-homomorphism φ : C0((0, 1], M2) −→ P∗∗, f ⊗ eij |−→ f(h)Eij , f ∈ C0((0, 1]).” · Show in paper

A point-norm continuous family of unital ∗*-homomorphisms θr:M2→M2⊗k\theta_r : M_2 \to M_2^{\otimes k} parametrized by the simplex Δk−1\Delta_{k-1} is constructed inductively using a unitary flip path U(s)U(s) to move M2M_2 continuously between tensor factors in the cube model C⊗kC^{\otimes k}.

Source excerpts

“Lemma 5.2. For every k ≥ 1, there is a point-norm continuous family of unital ∗-homomorphisms θr : M2 −→ M⊗k2 , r ∈ Δk−1 = { r ∈ [0, 1]k : ∑i ri = 1 } , whose range uses only the tensor factors indexed by {i : ri > 0}. At the j-th vertex, θr is the j-th coordinate inclusion.” · Show in paper

The Rørdam–Winter dimension-drop criterion is proved without stable rank one by combining Gram-type order-zero maps, a transported defect homomorphism φ:Her(a)→B\varphi : \mathrm{Her}(a) \to B, and a weighted flip unitary path U(t)U(t) on the overlap.

Source excerpts

“The flip where the two pieces overlap. Let F ∈ M2 ⊗ M2 be the flip unitary. Choose a continuous function θ : [0, 1] → [0, 1] that is zero on [0, 1/3] and one on [2/3, 1], and set U(t) = 1 + F 2 + eiπθ(t) 1 − F 2 .” · Show in paper

Separable reduction of a characterless unital C∗C^*-algebra BB is achieved by taking the C∗C^*-subalgebra generated by 1B1_B and the finitely many elements appearing in a commutator sum ∑i=1rhi[bi,ci]ki\sum_{i=1}^r h_i [b_i, c_i] k_i within norm distance 11 of 1B1_B.

Source excerpts

“Lemma 6.1. Every nonzero unital C∗-algebra B without characters contains a separable C∗-subalgebra D without characters and with 1D = 1B.” · Show in paper