# RAIL: Risk-Averse Imitation Learning

arXiv: https://arxiv.org/abs/1707.06658 · HTML: https://mlsanity.com/paper/1707.06658/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Jul 2017.

## Built on by (most cited first, 9 of 9)

- [Pre-Trained Language Models for Interactive Decision-Making](https://mlsanity.com/paper/2202.01771/) (2022)
- [Risk-Averse Offline Reinforcement Learning](https://mlsanity.com/paper/2102.05371/) (2021)
- [Simulating Emergent Properties of Human Driving Behavior Using Multi-Agent Reward Augmented Imitation Learning](https://mlsanity.com/paper/1903.05766/) (2019)
- [Risk-Sensitive Generative Adversarial Imitation Learning](https://mlsanity.com/paper/1808.04468/) (2018)
- [TW-CRL: Time-Weighted Contrastive Reward Learning for Efficient Inverse Reinforcement Learning](https://mlsanity.com/paper/2504.05585/) (2025)
- [Efficient Open-world Reinforcement Learning via Knowledge Distillation and Autonomous Rule Discovery](https://mlsanity.com/paper/2311.14270/) (2023)
- [Don't do it: Safer Reinforcement Learning With Rule-based Guidance](https://mlsanity.com/paper/2212.13819/) (2022)
- [Generalizable Image Repair for Robust Visual Control](https://mlsanity.com/paper/2503.05911/) (2025)
- [MoMo: Conditioned Contrastive Representation Learning for Preference-Modulated Planning](https://mlsanity.com/paper/2605.08512/) (2026)

## Abstract (arXiv)

Imitation learning algorithms learn viable policies by imitating an expert's behavior when reward signals are not available. Generative Adversarial Imitation Learning (GAIL) is a state-of-the-art algorithm for learning policies when the expert's behavior is available as a fixed set of trajectories. We evaluate in terms of the expert's cost function and observe that the distribution of trajectory-costs is often more heavy-tailed for GAIL-agents than the expert at a number of benchmark continuous-control tasks. Thus, high-cost trajectories, corresponding to tail-end events of catastrophic failure, are more likely to be encountered by the GAIL-agents than the expert. This makes the reliability of GAIL-agents questionable when it comes to deployment in risk-sensitive applications like robotic surgery and autonomous driving. In this work, we aim to minimize the occurrence of tail-end events by minimizing tail risk within the GAIL framework. We quantify tail risk by the Conditional-Value-at-Risk (CVaR) of trajectories and develop the Risk-Averse Imitation Learning (RAIL) algorithm. We observe that the policies learned with RAIL show lower tail-end risk than those of vanilla GAIL. Thus the proposed RAIL algorithm appears as a potent alternative to GAIL for improved reliability in risk-sensitive applications.