Online influence maximization under independent cascade model with semi-bandit feedback

Wen, Zheng; Kveton, Branislav; Valko, Michal; Vaswani, Sharan

Type de document :

Communication dans un congrès avec actes

Titre :

Online influence maximization under independent cascade model with semi-bandit feedback

Auteur(s) :

Wen, Zheng [Auteur]
Adobe Research
Kveton, Branislav [Auteur]
Adobe Research
Valko, Michal [Auteur]

Sequential Learning [SEQUEL]
Vaswani, Sharan [Auteur]
University of British Columbia [Canada] [UBC]

Titre de la manifestation scientifique :

Neural Information Processing Systems

Ville :

Long Beach

Pays :

Etats-Unis d'Amérique

Date de début de la manifestation scientifique :

2017-12-04

Discipline(s) HAL :

Statistiques [stat]/Machine Learning [stat.ML]

Résumé en anglais : [en]

We study the online influence maximization problem in social networks under the independent cascade model. Specifically, we aim to learn the set of " best influencers " in a social network online while repeatedly interacting ...
Lire la suite >We study the online influence maximization problem in social networks under the independent cascade model. Specifically, we aim to learn the set of " best influencers " in a social network online while repeatedly interacting with it. We address the challenges of (i) combinatorial action space, since the number of feasible influencer sets grows exponentially with the maximum number of influencers, and (ii) limited feedback, since only the influenced portion of the network is observed. Under a stochastic semi-bandit feedback, we propose and analyze IMLinUCB, a computationally efficient UCB-based algorithm. Our bounds on the cumulative regret are polynomial in all quantities of interest, achieve near-optimal dependence on the number of interactions and reflect the topology of the network and the activation probabilities of its edges, thereby giving insights on the problem complexity. To the best of our knowledge, these are the first such results. Our experiments show that in several representative graph topologies, the regret of IMLinUCB scales as suggested by our upper bounds. IMLinUCB permits linear generalization and thus is both statistically and computationally suitable for large-scale problems. Our experiments also show that IMLinUCB with linear generalization can lead to low regret in real-world online influence maximization.Lire moins >

Langue :

Anglais

Comité de lecture :

Oui

Audience :

Internationale

Vulgarisation :

Non

Projet ANR :

Inférence bayésienne à ressources limitées - données massives et modèles coûteux

Collections :