For any abelian variety J over a global field k and an isogeny phi: J -> J,
the Selmer group Sel^phi(J,k) is a subgroup of the Galois cohomology group
H^1(Gal(k^s/k),J[phi]), defined in terms of local data. When J is the Jacobian
of a cyclic cover of P^1, the Selmer group has a quotient by a subgroup of
order at most 2 that is isomorphic to the `fake Selmer group', whose definition
is more amenable to explicit computations.
These notes represent an extended version of a talk I gave for the
participants of the IMO 2009 and other interested people. We introduce
diophantine equations and show evidence that it can be hard to solve them. Then
we demonstrate how one can solve a specific equation related to numbers
occurring several times in Pascal's Triangle with state-of-the-art methods.
We answer a question asked by Hajdu and Tengely: The only arithmetic
progression in coprime integers of the form (a^2, b^2, c^2, d^5) is (1, 1, 1,
1).
In this paper, we generalise results obtained earlier by John Cremona and the
author on the reduction theory of binary forms, which describe positive
zero-cycles in P^1, to positive zero-cycles (or point clusters) in projective
spaces of arbitrary dimension. This should have applications to more general
projective varieties in P^n, by associating a suitable positive zero-cycle to
them in an PGL(n+1)-invariant way. We discuss this in the case of (smooth)
plane curves.
In this paper, we generalise results obtained earlier by John Cremona and the
author on the reduction theory of binary forms, which describe positive
zero-cycles in P^1, to positive zero-cycles (or point clusters) in projective
spaces of arbitrary dimension. This should have applications to more general
projective varieties in P^n, by associating a suitable positive zero-cycle to
them in an PGL(n+1)-invariant way. We discuss this in the case of (smooth)
plane curves.