On Bits and Bandits: Quantifying the Regret-Information Trade-off

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shufaro, Itai, Merlis, Nadav, Weinberger, Nir, Mannor, Shie
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915165717397504
author Shufaro, Itai
Merlis, Nadav
Weinberger, Nir
Mannor, Shie
author_facet Shufaro, Itai
Merlis, Nadav
Weinberger, Nir
Mannor, Shie
contents In many sequential decision problems, an agent performs a repeated task. He then suffers regret and obtains information that he may use in the following rounds. However, sometimes the agent may also obtain information and avoid suffering regret by querying external sources. We study the trade-off between the information an agent accumulates and the regret it suffers. We invoke information-theoretic methods for obtaining regret lower bounds, that also allow us to easily re-derive several known lower bounds. We introduce the first Bayesian regret lower bounds that depend on the information an agent accumulates. We also prove regret upper bounds using the amount of information the agent accumulates. These bounds show that information measured in bits, can be traded off for regret, measured in reward. Finally, we demonstrate the utility of these bounds in improving the performance of a question-answering task with large language models, allowing us to obtain valuable insights.
format Preprint
id arxiv_https___arxiv_org_abs_2405_16581
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle On Bits and Bandits: Quantifying the Regret-Information Trade-off
Shufaro, Itai
Merlis, Nadav
Weinberger, Nir
Mannor, Shie
Machine Learning
In many sequential decision problems, an agent performs a repeated task. He then suffers regret and obtains information that he may use in the following rounds. However, sometimes the agent may also obtain information and avoid suffering regret by querying external sources. We study the trade-off between the information an agent accumulates and the regret it suffers. We invoke information-theoretic methods for obtaining regret lower bounds, that also allow us to easily re-derive several known lower bounds. We introduce the first Bayesian regret lower bounds that depend on the information an agent accumulates. We also prove regret upper bounds using the amount of information the agent accumulates. These bounds show that information measured in bits, can be traded off for regret, measured in reward. Finally, we demonstrate the utility of these bounds in improving the performance of a question-answering task with large language models, allowing us to obtain valuable insights.
title On Bits and Bandits: Quantifying the Regret-Information Trade-off
topic Machine Learning
url https://arxiv.org/abs/2405.16581