Factorization Bandits for Interactive Recommendation

By Qingyun Wu, Hongning Wang. 2017

122 citations recorded by the source corpus.

We perform online interactive recommendation via a factorization-based bandit algorithm. Low-rank matrix completion is performed over an incrementally constructed user-item preference matrix, where an upper confidence bound based item selection strategy is developed to balance the exploit/explore trade-off during online learning. Observable contextual features and dependency among users (e.g., social influence) are leveraged to improve the algorithm's convergence rate and help conquer cold-start in recommendation. A high probability sublinear upper regret bound is proved for the developed algorithm, where considerable regret reduction is achieved on both user and item sides. Extensive experimentations on both simulations and large-scale real-world datasets confirmed the advantages of the proposed algorithm compared with several state-of-the-art factorization-based and bandit-based collaborative filtering methods.

Subfields: Electrical and Electronic Engineering, Information Systems, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Electrical and Electronic Engineering

In Information Systems

In Management Science and Operations Research

Citations in the corpus

DRN: A Deep Reinforcement Learning Framework for News Recommendation (2018, methodology): “The most straightforward strategies to do exploration in reinforcement learning are ϵ - greedy [31] and UCB [23]. ϵ - greedy will randomly recommend new items with a probability of ϵ , while UCB will pick items that have not been explored for many times (because these items may have larger variance)”
Estimation-Action-Reflection: Towards Deep Interaction Between Conversational and Recommender Systems (2020, background): “ferences. This intrinsic limitation motivates online recommendation. Its target is to adapt the recommendation results with the user’s online actions [25]. Many model it as a multi-arm bandit problem [34, 36, 37] , strategically demonstrating items to users for useful feedback. [39] makes the prelim”
Bias and Debias in Recommender System: A Survey and Future Directions (2020, background): “te the agent. Therefore, RL could balance the competition between the exploitation and exploration and maximize each user’s long term satisfaction with the system [13]. Some recent work [144], [145], [146], [147] balance exploitation and exploration in bandit setting with "-greedy, Boltzmann Ex”
Artificial intelligence in recommender systems (2020, background): “There is also some work which treats the interactive process between the user and the recommender system as a multi-arm bandit problem [145] and later extended with contextual information [146, 147].”
Meta-Adaptive Nonlinear Control: Theory and Algorithms (2021, methodology): “Outside the context of control, such bilinear models are also studied in representation learning [4, 5] and factorization bandits [29, 30].”
Advances and Challenges in Conversational Recommender Systems: A Survey (2021, methodology): “Some studies propose interactive recommender systems [57, 176, 20, 225], which mainly focus on improving the recommendation strategy online by leveraging real-time user feedback on previous recommended items.”