On Arthur's multiplicity formula for G₂ [NT seminar]

The Langlands philosophy predicts that automorphic forms should be related to Galois representations.
In this spirit, Arthur has outlined a conjectural description of discrete automorphic forms in terms of
"A-parameters". For the split group G₂, this description was proved for the most degenerate A-parameters
by Gan, Gurevich, and Jiang. We present a proof, in certain cases, for the least degenerate (but still
non-tempered!) A-parameters. As in work of Gan, Gurevich, and Jiang, our proof relies on exceptional
theta correspondences.

This is joint with P. Bakić, A. Horawa, and N. Sweeting.