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

Two-step monodromy of special quasi-projective varieties

Let VV be a connected smooth complex quasi-projective variety and let a:V→A=Alb(V)a : V \to A = \mathrm{Alb}(V) be its algebraic quasi-Albanese map. Suppose aa is dominant, choose a smooth connected general fibre FF and a point v∈Fv \in F, and assume that the sequence of ordinary fundamental groups π1(F,v)→Γ:=π1(V,v)→a∗Λ:=π1(A,a(v))→1\pi_1(F, v) \to \Gamma := \pi_1(V, v) \xrightarrow{a_*} \Lambda := \pi_1(A, a(v)) \to 1 is exact. Then every canonical rational unipotent quotient of Γ\Gamma has nilpotency class at most two, the full rational unipotent completion is a finite-dimensional unipotent algebraic group of class at most two, and every finite-dimensional complex unipotent representation of Γ\Gamma also has image of class at most two.

Let VV be a connected smooth special complex quasi-projective variety. For every positive integer dd and every representation ρ:π1(V)→GLd(C)\rho : \pi_1(V) \to \mathrm{GL}_d(\mathbf{C}), the image ρ(π1(V))\rho(\pi_1(V)) has a subgroup of finite index and nilpotency class at most two.

Let VV be a connected smooth complex quasi-projective variety with logarithmic Kodaira dimension κˉ(V)=0\bar{\kappa}(V) = 0, and put Γ=π1(V)\Gamma = \pi_1(V). For every positive integer dd and every representation ρ:Γ→GLd(C)\rho : \Gamma \to \mathrm{GL}_d(\mathbf{C}), the image ρ(Γ)\rho(\Gamma) has a subgroup of finite index and nilpotency class at most two, the full rational unipotent completion of Γ\Gamma is a finite-dimensional unipotent algebraic group of class at most two, and every finite-dimensional complex unipotent representation of Γ\Gamma also has image of class at most two.

Let EE be a complex elliptic curve, let q1,…,qnq_1, \dots, q_n be distinct points of P1\mathbf{P}^1 where n≥3n \ge 3, choose points e1,…,en∈Ee_1, \dots, e_n \in E, let b:Y=Bl{(ei,qi)}i=1n(E×P1)→E×P1b : Y = \mathrm{Bl}_{\{(e_i, q_i)\}_{i=1}^n}(E \times \mathbf{P}^1) \to E \times \mathbf{P}^1, write DiD_i for the strict transform of E×{qi}E \times \{q_i\}, and put D=∑iDiD = \sum_i D_i and X=Y∖DX = Y \setminus D. Then: (i) (Y,D)(Y, D) is a smooth projective special pair in the convention of [8, Definitions 2.1–2.9], including every upper model in that definition; (ii) a=(prE∘b)∣X:X→Ea = (\mathrm{pr}_E \circ b)|_X : X \to E is the algebraic quasi-Albanese map (after choosing origins), every fibre is connected, and its general fibre is F≃P1∖{q1,…,qn}F \simeq \mathbf{P}^1 \setminus \{q_1, \dots, q_n\} (which is not special); and (iii) a∗:π1(X)→π1(E)a_* : \pi_1(X) \to \pi_1(E) is an isomorphism, so π1(X)≃Z2\pi_1(X) \simeq \mathbf{Z}^2 and π1(F)→π1(X)\pi_1(F) \to \pi_1(X) is trivial.

The proof of [8, Claim 7.17] fails both numerically and in its stated justification when instantiated on the surface X=Y∖DX = Y \setminus D of Theorem 1.3, because it uses flatness of (Y×JJ′)main→Xˉ′(Y \times_J J')_{\mathrm{main}} \to \bar{X}' (which in this instance is the non-flat blowdown b:Y→Sb : Y \to S) although the preceding construction only arranged flatness over J′J', and consequently treats bb-exceptional divisors AiA_i as having boundary multiplicity ∞\infty instead of 11.

Methods

Equipping the canonical c-step nilpotent Lie quotients g_c of pi_1(V, v) with Hain's pointed mixed Hodge structures and using exactness of the quasi-Albanese fundamental-group sequence to bound the weight of the commutator abelianization B_c = k_c/[k_c, k_c].

Source excerpts

“We use Hain’s pointed construction [13, Theorem 6.3.1, p. 321] for this functoriality. It equips the completed Lie algebra with a pro-mixed Hodge structure compatible with its bracket.” · Show in paper

“Strictness of (2.7) gives W−3Bc = 0. On the other hand, strictness for the image kc ⊆ gc and (2.6) give W−2Bc = Bc. Therefore Bc is pure of weight − 2.” · Show in paper

Passing to a connected finite étale cover V_0 -> V corresponding to rho^{-1}(L^0), where L^0 = U x T is the identity component of the Zariski closure of rho(pi_1(V)), to apply the unipotent completion bound directly on V_0.

Source excerpts

“The third input makes V0 special. Apply the second input to the quasi-Albanese map of V0 itself: it is dominant, its general fibre is smooth and connected, and its ordinary fundamental-group sequence is exact. All the hypotheses of Theorem 1.2 therefore hold for V0.” · Show in paper

Constructing a special open surface X = Y \ D by blowing up E x P^1 at n >= 3 points (e_i, q_i) and deleting only the strict transforms D_i of E x {q_i} while retaining the exceptional affine lines A_i^o, and verifying specialness on all upper models via logarithmic pluricanonical section bounds and orbifold Riemann-Hurwitz on elliptic fibres.

Source excerpts

“The construction retains exceptional affine lines after deleting the strict transforms of horizontal elliptic curves.” · Show in paper

“For surface bases we use the growth of logarithmic pluricanonical sections. For curve bases we use the complete elliptic fibres of j and its retained multiplicity-one components.” · Show in paper