Safe Reinforcement Learning Using Advantage-Based Intervention Article Swipe

MAKE IMPACT

Nolan Wagener , Byron Boots , Cheng Chang · YOU? · · 2021 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2106.09110

Many sequential decision problems involve finding a policy that maximizes total reward while obeying safety constraints. Although much recent research has focused on the development of safe reinforcement learning (RL) algorithms that produce a safe policy after training, ensuring safety during training as well remains an open problem. A fundamental challenge is performing exploration while still satisfying constraints in an unknown Markov decision process (MDP). In this work, we address this problem for the chance-constrained setting. We propose a new algorithm, SAILR, that uses an intervention mechanism based on advantage functions to keep the agent safe throughout training and optimizes the agent's policy using off-the-shelf RL algorithms designed for unconstrained MDPs. Our method comes with strong guarantees on safety during both training and deployment (i.e., after training and without the intervention mechanism) and policy performance compared to the optimal safety-constrained policy. In our experiments, we show that SAILR violates constraints far less during training than standard safe RL and constrained MDP approaches and converges to a well-performing policy that can be deployed safely without intervention. Our code is available at https://github.com/nolanwagener/safe_rl.

Type

preprint

Language

en

Landing Page

http://arxiv.org/abs/2106.09110

PDF

https://arxiv.org/pdf/2106.09110

OA Status

green

Cited By

5

Related Works

10

OpenAlex ID

https://openalex.org/W3168575058

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W3168575058

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2106.09110

Digital Object Identifier

Title

Safe Reinforcement Learning Using Advantage-Based Intervention

Work title

Type

preprint

OpenAlex work type

Language

en

Primary language

Publication year

2021

Year of publication

Publication date

2021-06-16

Full publication date if available

Authors

Nolan Wagener, Byron Boots, Cheng Chang

List of authors in order

Landing page

https://arxiv.org/abs/2106.09110

Publisher landing page

PDF URL

https://arxiv.org/pdf/2106.09110

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2106.09110

Direct OA link when available

Concepts

Reinforcement learning, Markov decision process, Computer science, Intervention (counseling), Software deployment, Mathematical optimization, Process (computing), Training (meteorology), Risk analysis (engineering), Markov process, Operations research, Artificial intelligence, Engineering, Business, Mathematics, Operating system, Meteorology, Statistics, Physics, Psychology, Psychiatry

Top concepts (fields/topics) attached by OpenAlex

Cited by

5

Total citation count in OpenAlex

Citations by year (recent)

2025: 1, 2024: 1, 2023: 2, 2022: 1

Per-year citation counts (last 5 years)

Related works (count)

10

Other works algorithmically related by OpenAlex

Full payload