Fast Slate Policy Optimization: Going Beyond Plackett-Luce

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Sakhi, Otmane, Rohde, David, Chopin, Nicolas
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909056430505984
author Sakhi, Otmane
Rohde, David
Chopin, Nicolas
author_facet Sakhi, Otmane
Rohde, David
Chopin, Nicolas
contents An increasingly important building block of large scale machine learning systems is based on returning slates; an ordered lists of items given a query. Applications of this technology include: search, information retrieval and recommender systems. When the action space is large, decision systems are restricted to a particular structure to complete online queries quickly. This paper addresses the optimization of these large scale decision systems given an arbitrary reward function. We cast this learning problem in a policy optimization framework and propose a new class of policies, born from a novel relaxation of decision functions. This results in a simple, yet efficient learning algorithm that scales to massive action spaces. We compare our method to the commonly adopted Plackett-Luce policy class and demonstrate the effectiveness of our approach on problems with action space sizes in the order of millions.
format Preprint
id arxiv_https___arxiv_org_abs_2308_01566
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Fast Slate Policy Optimization: Going Beyond Plackett-Luce
Sakhi, Otmane
Rohde, David
Chopin, Nicolas
Machine Learning
Information Retrieval
An increasingly important building block of large scale machine learning systems is based on returning slates; an ordered lists of items given a query. Applications of this technology include: search, information retrieval and recommender systems. When the action space is large, decision systems are restricted to a particular structure to complete online queries quickly. This paper addresses the optimization of these large scale decision systems given an arbitrary reward function. We cast this learning problem in a policy optimization framework and propose a new class of policies, born from a novel relaxation of decision functions. This results in a simple, yet efficient learning algorithm that scales to massive action spaces. We compare our method to the commonly adopted Plackett-Luce policy class and demonstrate the effectiveness of our approach on problems with action space sizes in the order of millions.
title Fast Slate Policy Optimization: Going Beyond Plackett-Luce
topic Machine Learning
Information Retrieval
url https://arxiv.org/abs/2308.01566