Riemannian gossip algorithms for decentralized matrix completion (Bamdev Mishra - Amazon Development Centre India)

07.10.2016 10:15

In this talk, we propose novel gossip algorithms for the low-rank decentralized matrix completion problem. The proposed approach is on the Riemannian Grassmann manifold that allows local matrix completion by different agents while achieving asymptotic consensus on the global low-rank factors. The resulting approach is scalable and parallelizable. Our numerical experiments show the good performance of the proposed algorithms on various benchmarks, e.g., the Netflix dataset.

Lieu

salle 624, att. horaire inhabituel, Séminaire d'analyse numérique

Organisé par

Section de mathématiques

Intervenants

Bamdev Mishra, Amazon Development Centre India

entrée libre

Classement

Catégorie: Séminaire

Mots clés: analyse numérique

Plus d'infos

bamdevmishra.com/