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

Failure of rational injectivity for maximal coarse assembly

There exist finite connected graphs (Xj)j≥1(X_j)_{j\ge 1} with a common finite bound on vertex degree, a coarse disjoint union X=⨆j≥1XjX = \bigsqcup_{j\ge 1} X_j, and an infinite-order class α∈KX1(X)\alpha \in \mathrm{KX}_1(X) such that μXmax⁡(α)=0\mu_X^{\max}(\alpha) = 0; consequently, μXmax⁡⊗Q\mu_X^{\max} \otimes \mathbb{Q} is not injective.

The graphs Bj,Yj,XjB_j, Y_j, X_j are finite and connected, with vertex degrees bounded independently of jj, their projections to BjB_j are 11-Lipschitz, and each end copy of BjB_j is isometrically embedded in XjX_j.

The class α∈KX1(X)\alpha \in \mathrm{KX}_1(X) defined in (3.5) has infinite order, and in particular α⊗1≠0\alpha \otimes 1 \neq 0 in KX1(X)⊗Q\mathrm{KX}_1(X) \otimes \mathbb{Q}.

For a fixed integer r≥3r \ge 3 and nonzero integer nn, n(t+,j,r)∗τj−n(t−,j,r)∗τj≠0n(t_{+,j,r})_*\tau_j - n(t_{-,j,r})_*\tau_j \neq 0 in K1(Pr(Xj))K_1(P_r(X_j)) for all sufficiently large jj.

For the classes β∈KX1(B)\beta \in \mathrm{KX}_1(B) and α=(i+)∗β−(i−)∗β∈KX1(X)\alpha = (i_+)_*\beta - (i_-)_*\beta \in \mathrm{KX}_1(X), μXmax⁡(α)=(i+,A)∗μBmax⁡(β)−(i−,A)∗μBmax⁡(β)\mu_X^{\max}(\alpha) = (i_{+,A})_*\mu_B^{\max}(\beta) - (i_{-,A})_*\mu_B^{\max}(\beta).

For the projection f:Y→Bf : Y \to B there is a ∗*-homomorphism Φ:AB→AY\Phi : A_B \to A_Y such that (fAΦ)∗=id(f_A \Phi)_* = \mathrm{id} on K∗(AB)K_*(A_B); in particular, (fA)∗:K1(AY)→K1(AB)(f_A)_* : K_1(A_Y) \to K_1(A_B) is surjective.

The endpoint maps induce the same homomorphism (i+,AfA)∗=(i−,AfA)∗:K1(AY)→K1(AX)(i_{+,A} f_A)_* = (i_{-,A} f_A)_* : K_1(A_Y) \to K_1(A_X).

Methods

Construct a coarse disjoint union X of finite connected graphs X_j of uniformly bounded degree by placing congruence quotients Qkj(t)Q_{k_j(t)} of SL_3(Z) over 3-dimensional discrete tori B_j, with a central collar of constant level 2j and gradual one-level reductions to trivial fibers at the two ends.

Source excerpts

“We construct a graph with two torus ends and a long central region of constant fiber size. The fiber quotient changes one congruence level at a time outside this region. These gradual changes keep the degrees bounded, while the long central region will isolate the topological obstruction at any fixed Rips scale.” · Show in paper

“At height t, place a copy of Yj,kj (t) . Connect successive slices over the same base vertex by the reduction map from the larger quotient to the smaller one, or by the identity when the levels agree. The resulting finite graph is Xj .” · Show in paper

Lift edge transports in the Rips complex P_r(Y_j) from congruence level 2j to 3j to show that the pullback of u_1 cup u_2 is divisible by 2^j in H^2(P_r(Y_j); Z), and use a uniform Rips dimension bound to convert line-bundle roots into divisibility in K^1.

Source excerpts

“We will construct a Z/mZ-valued 1-cochain c with δc = −a ⌣ b by comparing edge transports after lifting from Q2j to Q3j .” · Show in paper

“The dimension bound converts the line-bundle roots into integral divisibility in K-theory.” · Show in paper

Use Kazhdan's property (T) for SL_3(Z) to obtain the projection P in M(A_Y) onto fiber constants, and combine it with horizontal matrix-entry transport L(T) to construct a *-homomorphism splitting Phi : A_B -> A_Y of f_A.

Source excerpts

“We will lift operators on B to operators on Y by transporting their matrix entries between these fibers and then projecting onto the constant vectors.” · Show in paper

“The continuous functional calculus thus gives a projection onto the invariant vectors in every representation, and (1 − ∆/2)^n converges to this projection in the maximal group norm. Denote its image in M(AY ) by P” · Show in paper

Place orthogonal copies of an operator at every height slice of X_j, rotate the stack continuously by one vertical step in M_2(A_X / K_X), and cancel the common interior copies modulo compact operators.

Source excerpts

“Instead, we place a copy of an operator at every height and compare this stack with its shift by one height. The common interior copies will cancel on K-theory.” · Show in paper