Making Sense of Reinforcement Learning and Probabilistic Inference Article Swipe

MAKE IMPACT

Brendan O’Donoghue , Ian Osband , Catalin Ionescu · YOU? · · 2020 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2001.00805

Reinforcement learning (RL) combines a control problem with statistical estimation: The system dynamics are not known to the agent, but can be learned through experience. A recent line of research casts `RL as inference' and suggests a particular framework to generalize the RL problem as probabilistic inference. Our paper surfaces a key shortcoming in that approach, and clarifies the sense in which RL can be coherently cast as an inference problem. In particular, an RL agent must consider the effects of its actions upon future rewards and observations: The exploration-exploitation tradeoff. In all but the most simple settings, the resulting inference is computationally intractable so that practical RL algorithms must resort to approximation. We demonstrate that the popular `RL as inference' approximation can perform poorly in even very basic problems. However, we show that with a small modification the framework does yield algorithms that can provably perform well, and we show that the resulting algorithm is equivalent to the recently proposed K-learning, which we further connect with Thompson sampling.

Type

article

Language

en

Landing Page

http://arxiv.org/abs/2001.00805

PDF

https://arxiv.org/pdf/2001.00805

OA Status

green

Cited By

12

References

46

Related Works

20

OpenAlex ID

https://openalex.org/W2996251520

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W2996251520

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2001.00805

Digital Object Identifier

Title

Making Sense of Reinforcement Learning and Probabilistic Inference

Work title

Type

article

OpenAlex work type

Language

en

Primary language

Publication year

2020

Year of publication

Publication date

2020-01-03

Full publication date if available

Authors

Brendan O’Donoghue, Ian Osband, Catalin Ionescu

List of authors in order

Landing page

https://arxiv.org/abs/2001.00805

Publisher landing page

PDF URL

https://arxiv.org/pdf/2001.00805

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/2001.00805

Direct OA link when available

Concepts

Reinforcement learning, Inference, Computer science, Probabilistic logic, Artificial intelligence, Machine learning, Simple (philosophy), Key (lock), Thompson sampling, Approximate inference, Statistical inference, Theoretical computer science, Bayesian probability, Mathematics, Computer security, Philosophy, Statistics, Epistemology

Top concepts (fields/topics) attached by OpenAlex

Cited by

12

Total citation count in OpenAlex

Citations by year (recent)

2023: 1, 2022: 2, 2021: 3, 2020: 5, 2018: 1

Per-year citation counts (last 5 years)

References (count)

46

Number of works referenced by this work

Related works (count)

20

Other works algorithmically related by OpenAlex

Full payload