# The Mirror Agent Model: a Bayesian Architecture for Interpretable Agent Behavior

arXiv: https://arxiv.org/abs/2609.05190 · HTML: https://mlsanity.com/paper/2609.05190/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 22 references that are in the map.
- First posted on arXiv 4 Sep 2026.

## Builds on (most cited first, 10 of 10)

- [OpenAI Gym](https://mlsanity.com/paper/1606.01540/) (2016)
- [Techniques for Interpretable Machine Learning](https://mlsanity.com/paper/1808.00033/) (2018)
- [Explainable Machine Learning for Scientific Insights and Discoveries](https://mlsanity.com/paper/1905.08883/) (2019)
- [Visualizing and Understanding Atari Agents](https://mlsanity.com/paper/1711.00138/) (2017)
- [Towards Interpretable Reinforcement Learning Using Attention Augmented Agents](https://mlsanity.com/paper/1906.02500/) (2019)
- [Contrastive Explanation: A Structural-Model Approach](https://mlsanity.com/paper/1811.03163/) (2018)
- [Exploratory Not Explanatory: Counterfactual Analysis of Saliency Maps for Deep Reinforcement Learning](https://mlsanity.com/paper/1912.05743/) (2019)
- [Explicability? Legibility? Predictability? Transparency? Privacy? Security? The Emerging Landscape of Interpretable Agent Behavior](https://mlsanity.com/paper/1811.09722/) (2018)
- [Towards Explainable AI Planning as a Service](https://mlsanity.com/paper/1908.05059/) (2019)
- [A Unifying Bayesian Formulation of Measures of Interpretability in Human-AI](https://mlsanity.com/paper/2104.10743/) (2021)

## Abstract (arXiv)

In this paper we illustrate a novel architecture generating interpretable behavior and explanations. We refer to this architecture as the Mirror Agent Model because it defines the observer model, that is the target of explicit and implicit communications, as a mirror of the agent's. With the goal of providing a general understanding of this work, we firstly show prior relevant results addressing the informative communication of agents intentions and the production of legible behavior. In the second part of the paper we furnish the architecture with novel capabilities for explanations through off-the-shelf saliency methods, followed by preliminary qualitative results.