# Dynamic Update-to-Data Ratio: Minimizing World Model Overfitting

arXiv: https://arxiv.org/abs/2303.10144 · HTML: https://mlsanity.com/paper/2303.10144/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Latent dynamics models and model-based RL, one of 11 directions in World models.
- Builds on 23 of its 35 references that are in the map.
- First posted on arXiv 17 Mar 2023.

## Placed in

- World models / [Latent dynamics models and model-based RL](https://mlsanity.com/wm/latent-dynamics-models-and-model-based-rl/)

## Builds on (most cited first, 23 of 23)

- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [Dream to Control: Learning Behaviors by Latent Imagination](https://mlsanity.com/paper/1912.01603/) (2019)
- [Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model](https://mlsanity.com/paper/1911.08265/) (2019)
- [Learning Latent Dynamics for Planning from Pixels](https://mlsanity.com/paper/1811.04551/) (2018)
- [Mastering Atari with Discrete World Models](https://mlsanity.com/paper/2010.02193/) (2020)
- [DeepMind Control Suite](https://mlsanity.com/paper/1801.00690/) (2018)
- [Deep Reinforcement Learning at the Edge of the Statistical Precipice](https://mlsanity.com/paper/2108.13264/) (2021)
- [Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models](https://mlsanity.com/paper/1805.12114/) (2018)
- [Model-Based Reinforcement Learning for Atari](https://mlsanity.com/paper/1903.00374/) (2019)
- [Reinforcement Learning with Augmented Data](https://mlsanity.com/paper/2004.14990/) (2020)
- [Data-Efficient Reinforcement Learning with Self-Predictive Representations](https://mlsanity.com/paper/2007.05929/) (2020)
- [Agent57: Outperforming the Atari Human Benchmark](https://mlsanity.com/paper/2003.13350/) (2020)
- [A Study on Overfitting in Deep Reinforcement Learning](https://mlsanity.com/paper/1804.06893/) (2018)
- [Assessing Generalization in Deep Reinforcement Learning](https://mlsanity.com/paper/1810.12282/) (2018)
- [Observational Overfitting in Reinforcement Learning](https://mlsanity.com/paper/1912.02975/) (2019)
- [A Dissection of Overfitting and Generalization in Continuous Reinforcement Learning](https://mlsanity.com/paper/1806.07937/) (2018)
- [Automated Reinforcement Learning (AutoRL): A Survey and Open Problems](https://mlsanity.com/paper/2201.03916/) (2022)
- [On the Importance of Hyperparameter Optimization for Model-based Reinforcement Learning](https://mlsanity.com/paper/2102.13651/) (2021)
- [Bayesian Optimization in AlphaGo](https://mlsanity.com/paper/1812.06855/) (2018)
- [MBRL-Lib: A Modular Library for Model-based Reinforcement Learning](https://mlsanity.com/paper/2104.10159/) (2021)
- [Adapting Behaviour for Learning Progress](https://mlsanity.com/paper/1912.06910/) (2019)
- [Mitigating Planner Overfitting in Model-Based Reinforcement Learning](https://mlsanity.com/paper/1812.01129/) (2018)
- [Adaptively Calibrated Critic Estimates for Deep Reinforcement Learning](https://mlsanity.com/paper/2111.12673/) (2021)

## Built on by (most cited first, 4 of 4)

- [Langevin Soft Actor-Critic: Efficient Exploration through Uncertainty-Driven Critic Learning](https://mlsanity.com/paper/2501.17827/) (2025)
- [Mind the Model, Not the Agent: The Primacy Bias in Model-based RL](https://mlsanity.com/paper/2310.15017/) (2023)
- [Measuring the Value of World-Model Updates: A Counterfactual Utility Protocol for Continual Adaptation](https://mlsanity.com/paper/2609.10954/) (2026)
- [Optimal Training-Time Scaling in Gradual Adaptation](https://mlsanity.com/paper/2608.04927/) (2026)

## Abstract (arXiv)

Early stopping based on the validation set performance is a popular approach to find the right balance between under- and overfitting in the context of supervised learning. However, in reinforcement learning, even for supervised sub-problems such as world model learning, early stopping is not applicable as the dataset is continually evolving. As a solution, we propose a new general method that dynamically adjusts the update to data (UTD) ratio during training based on under- and overfitting detection on a small subset of the continuously collected experience not used for training. We apply our method to DreamerV2, a state-of-the-art model-based reinforcement learning algorithm, and evaluate it on the DeepMind Control Suite and the Atari 100k benchmark. The results demonstrate that one can better balance under- and overestimation by adjusting the UTD ratio with our approach compared to the default setting in DreamerV2 and that it is competitive with an extensive hyperparameter search which is not feasible for many applications. Our method eliminates the need to set the UTD hyperparameter by hand and even leads to a higher robustness with regard to other learning-related hyperparameters further reducing the amount of necessary tuning.