---
title: "Toward Optimal Solution for the Context-Attentive Bandit Problem"
authors:
  - "Djallel Bouneffouf"
  - "Raphael Feraud"
  - "Sohini Upadhyay"
  - "Irina Rish"
  - "Yasaman Khazaeni"
doi: "10.24963/ijcai.2021/481"
url: "https://ijcai.org/proceedings/2021/481"
pdf: "https://ijcai.org/proceedings/2021/0481.pdf"
venue: "Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence"
year: 2021
pages: "3493-3500"
track: "Main Track"
keywords:
  - "Data Mining: Recommender Systems"
  - "Machine Learning Applications: Applications of Reinforcement Learning"
  - "Data Mining: Big Data, Large-Scale Systems"
---

# Toward Optimal Solution for the Context-Attentive Bandit Problem

Djallel Bouneffouf, Raphael Feraud, Sohini Upadhyay, Irina Rish, Yasaman Khazaeni

*Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence*, pages 3493-3500. DOI: [10.24963/ijcai.2021/481](https://doi.org/10.24963/ijcai.2021/481)

## Abstract

In various recommender system applications, from medical diagnosis to dialog systems, due to observation costs only a small subset of a potentially large number of context variables can be observed at each iteration; however, the agent has a freedom to choose which variables to observe.  In this paper, we analyze and  extend an online learning framework known as Context-Attentive Bandit, We derive a novel algorithm, called Context-Attentive Thompson Sampling (CATS), which builds upon the Linear Thompson Sampling approach,  adapting it to Context-Attentive Bandit  setting.  We provide a theoretical regret analysis and an extensive empirical evaluation demonstrating  advantages of the proposed approach over several baseline methods  on a variety of real-life datasets.