Abstract:Let $A$ be an abelian variety over the function field $K$ of a compact Riemann surface $B$. Fix a model $f \colon \mathcal{A} \to B$ of $A/K$ and an effective horizontal divisor $\mathcal{D} \subset \mathcal{A}$. We study $(S, \mathcal{D})$-integral sections $\sigma$ of $\mathcal{A}$ where $S \subset B$ is arbitrary. These sections $\sigma$ are algebraic and satisfy the geometric condition $f(\sigma(B) \cap \mathcal{D})\subset S$. Developing the idea of Parshin, we formulate a hyperbolic-homotopic height of such sections as a substitute for intersection theory to establish new results concerning the finiteness and the polynomial growth of large unions of $(S, \mathcal{D})$-integral points where $S$ is only required to be finite in a thin analytic open subset of $B$. Such results are out of reach of purely algebraic methods and imply new evidence and interesting phenomena to the Geometric Lang-Vojta conjecture.
What problem does this paper attempt to address?
The problems that this paper attempts to solve mainly focus on generalized integral points on abelian varieties and evidence and new phenomena of the geometric Lang - Vojta conjecture. Specifically, the author studied the properties of $(S, D)$ - integral sections $\sigma$ in an abelian variety family $f: A \to B$ over a compact Riemann surface, given an effective horizontal divisor $D\subset A$ and an arbitrary set $S\subset B$.
### Core problems of the paper
1. **Finiteness and polynomial growth of integral sections**:
- The author explored the number and distribution of $(S, D)$ - integral sections, especially when $S$ can vary and is only finite in a thin analytic open subset of $B$.
- These results go beyond the scope of pure algebraic methods and provide new evidence and interesting phenomena for the geometric Lang - Vojta conjecture.
2. **Applications of the geometric Lang - Vojta conjecture**:
- A weak form of the geometric Lang - Vojta conjecture states that, under appropriate conditions, there exists a proper closed subset $Z\subset X$ and a constant $m > 0$ such that for every section $\sigma$ not contained in $Z$, its corresponding extension $\bar{\sigma}$ satisfies the degree inequality:
\[
\deg_{\bar{C}} \bar{\sigma}^*D \leq m(2g(\bar{C}) - 2+\# S)
\]
- The author provided new evidence for this conjecture by studying integral sections on abelian variety families, especially in the case where $\bar{X}\to\bar{C}$ is an abelian variety family over a compact Riemann surface.
3. **Hyperbolic length estimates**:
- The author introduced a new technique, the hyperbolic - homotopic method, to replace the classical intersection theory.
- In particular, the author proved a linear bound on the hyperbolic length of loops on a compact Riemann surface, which plays a crucial role in dealing with quantitative problems of integral sections.
### Main results
- **Theorem A**: Gives a polynomial upper bound on the number of integral sections, in the specific form:
\[
\# I_s \mod \text{Tr}_{K/C}(A)(C) \leq m(s + 1)^{2\dim A\cdot \text{rank}\pi_1(B_0)}
\]
where $I_s$ represents the set of integral sections satisfying specific conditions.
- **Theorem B**: Provides a linear estimate on the hyperbolic length of loops on a compact Riemann surface, which is of great significance for understanding the distribution of integral sections.
- **Corollary 1.3**: Under certain conditions, for every section $\sigma$, its corresponding degree satisfies:
\[
\deg_B \sigma^*D < M
\]
These results not only extend the existing theories but also provide new perspectives and evidence for the geometric Lang - Vojta conjecture.