MASCOT: Towards Multi-Agent Socio-Collaborative Companion Systems

TL;DR

MASCOT is a framework that enhances multi-agent systems' ability to maintain individual personas and produce diverse, constructive social interactions, addressing key issues like persona collapse and social sycophancy.

Contribution

It introduces a bi-level optimization strategy combining persona fidelity and collaborative dialogue quality, advancing the development of socially intelligent multi-agent systems.

Findings

Significant improvements in Persona Consistency (+14.1)

Enhanced Social Contribution (+10.6)

Outperforms state-of-the-art baselines in multiple domains

Abstract

Multi-agent systems (MAS) have recently emerged as promising socio-collaborative companions for emotional and cognitive support. However, these systems frequently suffer from persona collapse--where agents revert to generic, homogenized assistant behaviors--and social sycophancy, which produces redundant, non-constructive dialogue. We propose MASCOT, a generalizable framework for multi-perspective socio-collaborative companions. MASCOT introduces a novel bi-level optimization strategy to harmonize individual and collective behaviors: 1) Persona-Aware Behavioral Alignment, an RLAIF-driven pipeline that finetunes individual agents for strict persona fidelity to prevent identity loss; and 2) Collaborative Dialogue Optimization, a meta-policy guided by group-level rewards to ensure diverse and productive discourse. Extensive evaluations across psychological support and workplace domains demonstrate that MASCOT significantly outperforms state-of-the-art baselines, achieving improvements of up to +14.1 in Persona Consistency and +10.6 in Social Contribution. Our framework provides a practical roadmap for engineering the next generation of socially intelligent multi-agent systems.