Résumé : In this talk we shall discuss our recent results on the Adams' conjecture on theta correspondence. In more words, given a representation of a classical group (in our case, symplectic or even orthogonal) belonging to a local Arthur packet, Adams predicts that, under certain assumptions, its theta lift (i.e. a corresponding irreducible representation of the other group in a dual reductive pair), provided it is non-zero, is also in A-packet which can be easily described in terms of the original one. Mœglin gave some partial results, specifically, in case when the original representation is square-integrable. We are able to extend her results to the case of so called Arthur packets with the discrete diagonal restriction. Moreover, it seems that Arthur packet encapsulates lot of additional information even in relation to theta correspondence, e.g. we can easily read of from it the first occurrence index for the given representation in it. Adams conjecture takes an unexpectedly elegant form for the representations in discrete diagonal restriction packets. Also, we are able to pinpoint exactly how low in theta towers we can go with this description of the theta lifts which belong to Arthur packets, we can also address some other related conjectures due to Mœglin. This is joint work with Petar Baki.

Mots-Clés : Arthur packets; theta correspondence; classical groups