# Safe Learning Under Irreversible Dynamics via Asking for Help

arXiv: https://arxiv.org/abs/2502.14043 · HTML: https://mlsanity.com/paper/2502.14043/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 75 references that are in the map.
- First posted on arXiv 19 Feb 2025.

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

- [Generalized Out-of-Distribution Detection: A Survey](https://mlsanity.com/paper/2110.11334/) (2021)
- [Minimax Regret Bounds for Reinforcement Learning](https://mlsanity.com/paper/1703.05449/) (2017)
- [An Algorithmic Perspective on Imitation Learning](https://mlsanity.com/paper/1811.06711/) (2018)
- [Inverse Reward Design](https://mlsanity.com/paper/1711.02827/) (2017)
- [An Overview of Catastrophic AI Risks](https://mlsanity.com/paper/2306.12001/) (2023)
- [Almost Optimal Model-Free Reinforcement Learning via Reference-Advantage Decomposition](https://mlsanity.com/paper/2004.10019/) (2020)
- [Learning Policies with Zero or Bounded Constraint Violation for Constrained MDPs](https://mlsanity.com/paper/2106.02684/) (2021)
- [What Doubling Tricks Can and Can't Do for Multi-Armed Bandits](https://mlsanity.com/paper/1803.06971/) (2018)
- [Smoothed Analysis with Adaptive Adversaries](https://mlsanity.com/paper/2102.08446/) (2021)
- [State-wise Safe Reinforcement Learning: A Survey](https://mlsanity.com/paper/2302.03122/) (2023)
- [Smoothed Online Learning is as Easy as Statistical Learning](https://mlsanity.com/paper/2202.04690/) (2022)
- [Nearly Minimax Optimal Reinforcement Learning for Discounted MDPs](https://mlsanity.com/paper/2010.00587/) (2020)
- [Provably Safe Reinforcement Learning: Conceptual Analysis, Survey, and Benchmarking](https://mlsanity.com/paper/2205.06750/) (2022)
- [Online Learning with Abstention](https://mlsanity.com/paper/1703.03478/) (2017)
- [TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI](https://mlsanity.com/paper/2306.06924/) (2023)
- [Regret Bounds for Discounted MDPs](https://mlsanity.com/paper/2002.05138/) (2020)
- [Online Regret Bounds for Undiscounted Continuous Reinforcement Learning](https://mlsanity.com/paper/1302.2550/) (2013)
- [Learning Adversarial MDPs with Stochastic Hard Constraints](https://mlsanity.com/paper/2403.03672/) (2024)
- [Enabling Pedestrian Safety using Computer Vision Techniques: A Case Study of the 2018 Uber Inc. Self-driving Car Crash](https://mlsanity.com/paper/1805.11815/) (2018)
- [Pessimism About Unknown Unknowns Inspires Conservatism](https://mlsanity.com/paper/2006.08753/) (2020)
- [Curiosity Killed or Incapacitated the Cat and the Asymptotically Optimal Agent](https://mlsanity.com/paper/2006.03357/) (2020)
- [Avoiding Catastrophe in Online Learning by Asking for Help](https://mlsanity.com/paper/2402.08062/) (2024)
- [Delegative Reinforcement Learning: learning to avoid traps with a little help](https://mlsanity.com/paper/1907.08461/) (2019)

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

- [Learning When Not to Learn: Risk-Sensitive Abstention in Bandits with Unbounded Rewards](https://mlsanity.com/paper/2510.14884/) (2025)

## Abstract (arXiv)

Most learning algorithms with formal regret guarantees essentially rely on trying all possible behaviors, which is problematic when some errors cannot be recovered from. Instead, we allow the learning agent to ask for help from a mentor and to transfer knowledge between similar states. We show that this combination enables the agent to learn both safely and effectively. Under standard online learning assumptions, we provide an algorithm whose regret and number of mentor queries are both sublinear in the time horizon for Markov decision processes with irreversible dynamics and infinite state spaces. Our proof involves a sequence of three reductions, making our result more general than a single algorithm. Conceptually, our result may be the first formal proof that it is possible for an agent to obtain high reward while becoming self-sufficient in an unknown, unbounded, and high-stakes environment without resets.