# Domain-Independent Optimistic Initialization for Reinforcement Learning

arXiv: https://arxiv.org/abs/1410.4604 · HTML: https://mlsanity.com/paper/1410.4604/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Oct 2014.

## Built on by (most cited first, 8 of 8)

- [Flipping Coins to Estimate Pseudocounts for Exploration in Reinforcement Learning](https://mlsanity.com/paper/2306.03186/) (2023)
- [An Empirical Study of Implicit Regularization in Deep Offline RL](https://mlsanity.com/paper/2207.02099/) (2022)
- [Optimistic Reinforcement Learning by Forward Kullback-Leibler Divergence Optimization](https://mlsanity.com/paper/2105.12991/) (2021)
- [Optimistic Temporal Difference Learning for 2048](https://mlsanity.com/paper/2111.11090/) (2021)
- [Cyclophobic Reinforcement Learning](https://mlsanity.com/paper/2308.15911/) (2023)
- [Advancing Weight and Channel Sparsification with Enhanced Saliency](https://mlsanity.com/paper/2502.03658/) (2025)
- [Quality-Aware Exploration Budget Allocation for Cooperative Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2605.01865/) (2026)
- [Space-sampled Value Decay: Forgetting Mechanisms for Non-stationary Reinforcement Learning](https://mlsanity.com/paper/2606.11797/) (2026)

## Abstract (arXiv)

In Reinforcement Learning (RL), it is common to use optimistic initialization of value functions to encourage exploration. However, such an approach generally depends on the domain, viz., the scale of the rewards must be known, and the feature representation must have a constant norm. We present a simple approach that performs optimistic initialization with less dependence on the domain.