# UCB Exploration via Q-Ensembles

arXiv: https://arxiv.org/abs/1706.01502 · HTML: https://mlsanity.com/paper/1706.01502/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Jun 2017.

## Built on by (most cited first, 1 of 1)

- [An Experimental Design Perspective on Model-Based Reinforcement Learning](https://mlsanity.com/paper/2112.05244/) (2021)

## Abstract (arXiv)

We show how an ensemble of Q^*-functions can be leveraged for more effective exploration in deep reinforcement learning. We build on well established algorithms from the bandit setting, and adapt them to the Q-learning setting. We propose an exploration strategy based on upper-confidence bounds (UCB). Our experiments show significant gains on the Atari benchmark.