# An Approach to Partial Observability in Games: Learning to Both Act and Observe

arXiv: https://arxiv.org/abs/2108.05701 · HTML: https://mlsanity.com/paper/2108.05701/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 18 references that are in the map.
- First posted on arXiv 11 Aug 2021.

## Builds on (most cited first, 16 of 16)

- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [Transformers in Vision: A Survey](https://mlsanity.com/paper/2101.01169/) (2021)
- [Learning Invariant Representations for Reinforcement Learning without Reconstruction](https://mlsanity.com/paper/2006.10742/) (2020)
- [Curriculum Learning for Reinforcement Learning Domains: A Framework and Survey](https://mlsanity.com/paper/2003.04960/) (2020)
- [Recurrent Models of Visual Attention](https://mlsanity.com/paper/1406.6247/) (2014)
- [Agent57: Outperforming the Atari Human Benchmark](https://mlsanity.com/paper/2003.13350/) (2020)
- [StarCraft II: A New Challenge for Reinforcement Learning](https://mlsanity.com/paper/1708.04782/) (2017)
- [Action Branching Architectures for Deep Reinforcement Learning](https://mlsanity.com/paper/1711.08946/) (2017)
- [Towards Interpretable Reinforcement Learning Using Attention Augmented Agents](https://mlsanity.com/paper/1906.02500/) (2019)
- [A Survey of Deep Reinforcement Learning in Video Games](https://mlsanity.com/paper/1912.10944/) (2019)
- [Exploratory Not Explanatory: Counterfactual Analysis of Saliency Maps for Deep Reinforcement Learning](https://mlsanity.com/paper/1912.05743/) (2019)
- [Neuroevolution of Self-Interpretable Agents](https://mlsanity.com/paper/2003.08165/) (2020)
- [Learning to Look Around: Intelligently Exploring Unseen Environments for Unknown Tasks](https://mlsanity.com/paper/1709.00507/) (2017)
- [Free-Lunch Saliency via Attention in Atari Agents](https://mlsanity.com/paper/1908.02511/) (2019)
- [Learning to predict where to look in interactive environments using deep recurrent q-learning](https://mlsanity.com/paper/1612.05753/) (2016)
- [Uncertainty Maximization in Partially Observable Domains: A Cognitive Perspective](https://mlsanity.com/paper/2102.11232/) (2021)

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

- [InfoChess: A Game of Adversarial Inference and a Laboratory for Quantifiable Information Control](https://mlsanity.com/paper/2604.15373/) (2026)

## Abstract (arXiv)

Reinforcement learning (RL) is successful at learning to play games where the entire environment is visible. However, RL approaches are challenged in complex games like Starcraft II and in real-world environments where the entire environment is not visible. In these more complex games with more limited visual information, agents must choose where to look and how to optimally use their limited visual information in order to succeed at the game. We verify that with a relatively simple model the agent can learn where to look in scenarios with a limited visual bandwidth. We develop a method for masking part of the environment in Atari games to force the RL agent to learn both where to look and how to play the game in order to study where the RL agent learns to look. In addition, we develop a neural network architecture and method for allowing the agent to choose where to look and what action to take in the Pong game. Further, we analyze the strategies the agent learns to better understand how the RL agent learns to play the game.