Saved in:
Bibliographic Details
Main Authors: Zamir, Ram, Rose, Kenneth
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.03054
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915006011932672
author Zamir, Ram
Rose, Kenneth
author_facet Zamir, Ram
Rose, Kenneth
contents The inherent trade-off in on-line learning is between exploration and exploitation. A good balance between these two (conflicting) goals can achieve a better long-term performance. Can we define an optimal balance? We propose to study this question through a backward-adaptive lossy compression system, which exhibits a "natural" trade-off between exploration and exploitation.
format Preprint
id arxiv_https___arxiv_org_abs_2411_03054
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Alternate Learning and Compression Approaching R(D)
Zamir, Ram
Rose, Kenneth
Information Theory
The inherent trade-off in on-line learning is between exploration and exploitation. A good balance between these two (conflicting) goals can achieve a better long-term performance. Can we define an optimal balance? We propose to study this question through a backward-adaptive lossy compression system, which exhibits a "natural" trade-off between exploration and exploitation.
title Alternate Learning and Compression Approaching R(D)
topic Information Theory
url https://arxiv.org/abs/2411.03054