The Search for Stability: Learning Dynamics of Strategic Publishers with Initial Documents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Madmon, Omer, Pipano, Idan, Reinman, Itamar, Tennenholtz, Moshe
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915417198428160
author Madmon, Omer
Pipano, Idan
Reinman, Itamar
Tennenholtz, Moshe
author_facet Madmon, Omer
Pipano, Idan
Reinman, Itamar
Tennenholtz, Moshe
contents We study a game-theoretic information retrieval model in which strategic publishers aim to maximize their chances of being ranked first by the search engine while maintaining the integrity of their original documents. We show that the commonly used Probability Ranking Principle (PRP) ranking scheme results in an unstable environment where games often fail to reach pure Nash equilibrium. We propose two families of ranking functions that do not adhere to the PRP principle. We provide both theoretical and empirical evidence that these methods lead to a stable search ecosystem, by providing positive results on the learning dynamics convergence. We also define the publishers' and users' welfare, demonstrate a possible publisher-user trade-off, and provide means for a search system designer to control it. Finally, we show how instability harms long-term users' welfare.
format Preprint
id arxiv_https___arxiv_org_abs_2305_16695
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle The Search for Stability: Learning Dynamics of Strategic Publishers with Initial Documents
Madmon, Omer
Pipano, Idan
Reinman, Itamar
Tennenholtz, Moshe
Computer Science and Game Theory
Information Retrieval
We study a game-theoretic information retrieval model in which strategic publishers aim to maximize their chances of being ranked first by the search engine while maintaining the integrity of their original documents. We show that the commonly used Probability Ranking Principle (PRP) ranking scheme results in an unstable environment where games often fail to reach pure Nash equilibrium. We propose two families of ranking functions that do not adhere to the PRP principle. We provide both theoretical and empirical evidence that these methods lead to a stable search ecosystem, by providing positive results on the learning dynamics convergence. We also define the publishers' and users' welfare, demonstrate a possible publisher-user trade-off, and provide means for a search system designer to control it. Finally, we show how instability harms long-term users' welfare.
title The Search for Stability: Learning Dynamics of Strategic Publishers with Initial Documents
topic Computer Science and Game Theory
Information Retrieval
url https://arxiv.org/abs/2305.16695