# (When) Are Contrastive Explanations of Reinforcement Learning Helpful?

arXiv: https://arxiv.org/abs/2211.07719 · HTML: https://mlsanity.com/paper/2211.07719/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 24 references that are in the map.
- First posted on arXiv 14 Nov 2022.

## Builds on (most cited first, 14 of 14)

- [Explanation in Artificial Intelligence: Insights from the Social Sciences](https://mlsanity.com/paper/1706.07269/) (2017)
- [Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negatives](https://mlsanity.com/paper/1802.07623/) (2018)
- [Verifiable Reinforcement Learning via Policy Extraction](https://mlsanity.com/paper/1805.08328/) (2018)
- [Explainable Reinforcement Learning Through a Causal Lens](https://mlsanity.com/paper/1905.10958/) (2019)
- [Establishing Appropriate Trust via Critical States](https://mlsanity.com/paper/1810.08174/) (2018)
- [Contrastive Explanations for Reinforcement Learning in terms of Expected Consequences](https://mlsanity.com/paper/1807.08706/) (2018)
- [Interpretability via Model Extraction](https://mlsanity.com/paper/1706.09773/) (2017)
- [Enabling Robots to Communicate their Objectives](https://mlsanity.com/paper/1702.03465/) (2017)
- [Contrastive Explanations with Local Foil Trees](https://mlsanity.com/paper/1806.07470/) (2018)
- [Local and Global Explanations of Agent Behavior: Integrating Strategy Summaries with Saliency Maps](https://mlsanity.com/paper/2005.08874/) (2020)
- [Exploring Computational User Models for Agent Policy Summarization](https://mlsanity.com/paper/1905.13271/) (2019)
- [Conservative Q-Improvement: Reinforcement Learning for an Interpretable Decision-Tree Policy](https://mlsanity.com/paper/1907.01180/) (2019)
- [Contrastive Explanations for Reinforcement Learning via Embedded Self Predictions](https://mlsanity.com/paper/2010.05180/) (2020)
- ["I Don't Think So": Summarizing Policy Disagreements for Agent Comparison](https://mlsanity.com/paper/2102.03064/) (2021)

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

- [Optimizing Heat Alert Issuance with Reinforcement Learning](https://mlsanity.com/paper/2312.14196/) (2023)

## Abstract (arXiv)

Global explanations of a reinforcement learning (RL) agent's expected behavior can make it safer to deploy. However, such explanations are often difficult to understand because of the complicated nature of many RL policies. Effective human explanations are often contrastive, referencing a known contrast (policy) to reduce redundancy. At the same time, these explanations also require the additional effort of referencing that contrast when evaluating an explanation. We conduct a user study to understand whether and when contrastive explanations might be preferable to complete explanations that do not require referencing a contrast. We find that complete explanations are generally more effective when they are the same size or smaller than a contrastive explanation of the same policy, and no worse when they are larger. This suggests that contrastive explanations are not sufficient to solve the problem of effectively explaining reinforcement learning policies, and require additional careful study for use in this context.