Toward Safe and Human-Aligned Game Conversational Recommendation via Multi-Agent Decomposition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hui, Zheng, Wei, Xiaokai, Jiang, Yexi, Gao, Kevin, Wang, Chen, Ong, Frank, Yoon, Se-eun, Pareek, Rachit, Gong, Michelle
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915814957907968
author Hui, Zheng
Wei, Xiaokai
Jiang, Yexi
Gao, Kevin
Wang, Chen
Ong, Frank
Yoon, Se-eun
Pareek, Rachit
Gong, Michelle
author_facet Hui, Zheng
Wei, Xiaokai
Jiang, Yexi
Gao, Kevin
Wang, Chen
Ong, Frank
Yoon, Se-eun
Pareek, Rachit
Gong, Michelle
contents Conversational recommender systems (CRS) have advanced with large language models, showing strong results in domains like movies. These domains typically involve fixed content and passive consumption, where user preferences can be matched by genre or theme. In contrast, games present distinct challenges: fast-evolving catalogs, interaction-driven preferences (e.g., skill level, mechanics, hardware), and increased risk of unsafe responses in open-ended conversation. We propose MATCHA, a multi-agent framework for CRS that assigns specialized agents for intent parsing, tool-augmented retrieval, multi-LLM ranking with reflection, explanation, and risk control which enabling finer personalization, long-tail coverage, and stronger safety. Evaluated on real user request dataset, MATCHA outperforms six baselines across eight metrics, improving Hit@5 by 20%, reducing popularity bias by 24%, and achieving 97.9% adversarial defense. Human and virtual-judge evaluations confirm improved explanation quality and user alignment.
format Preprint
id arxiv_https___arxiv_org_abs_2504_20094
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Toward Safe and Human-Aligned Game Conversational Recommendation via Multi-Agent Decomposition
Hui, Zheng
Wei, Xiaokai
Jiang, Yexi
Gao, Kevin
Wang, Chen
Ong, Frank
Yoon, Se-eun
Pareek, Rachit
Gong, Michelle
Information Retrieval
Computation and Language
Human-Computer Interaction
Conversational recommender systems (CRS) have advanced with large language models, showing strong results in domains like movies. These domains typically involve fixed content and passive consumption, where user preferences can be matched by genre or theme. In contrast, games present distinct challenges: fast-evolving catalogs, interaction-driven preferences (e.g., skill level, mechanics, hardware), and increased risk of unsafe responses in open-ended conversation. We propose MATCHA, a multi-agent framework for CRS that assigns specialized agents for intent parsing, tool-augmented retrieval, multi-LLM ranking with reflection, explanation, and risk control which enabling finer personalization, long-tail coverage, and stronger safety. Evaluated on real user request dataset, MATCHA outperforms six baselines across eight metrics, improving Hit@5 by 20%, reducing popularity bias by 24%, and achieving 97.9% adversarial defense. Human and virtual-judge evaluations confirm improved explanation quality and user alignment.
title Toward Safe and Human-Aligned Game Conversational Recommendation via Multi-Agent Decomposition
topic Information Retrieval
Computation and Language
Human-Computer Interaction
url https://arxiv.org/abs/2504.20094