Algebraic independence and diophantine approximation

点击翻译按钮获取中文摘要

- Main topic: A fascinating conjecture from transcendental number theory asserts that Q-linearly independent logarithms of algebraic numbers are algebraically independent over Q. Its solution would for example imply a conjecture of Leopoldt on the non-vanishing of the p-adic regulator of a number field for any prime number p. My main goal is to prove that any sufficiently large set of Q-linearly independent logarithms of algebraic numbers contains two algebraically independent elements. Some years ago, I showed that, in principle, well-known constructions are sufficient to capture all the needed information, in the form of a sequence of polynomials taking small values on large sets of points from a finitely generated subgroup of the algebraic group Ga x Gm. This requires new results that would encompass both Philippon's criterion for algebraic independence and the actual zero estimates. Recently, I was able to prove non-trivial estimates of this form for the one dimensional groups Ga and Gm, and I have partial results for the two-dimensional group Ga x Gm which I intend to develop.