MASCOT: Towards Multi-Agent Socio-Collaborative Companion Systems

Jul 1, 2026·
Yiyang Wang
Yiqiao Jin
Yiqiao Jin
,
Alex Cabral
,
Josiah Hester
· 1 min read
Abstract
Companion systems for everyday users must balance helpfulness, trust, and social context. We introduce MASCOT, a multi-agent socio-collaborative companion framework that coordinates specialized agents to support socially grounded, trustworthy interaction. MASCOT explicitly models social context, user goals, and inter-agent collaboration patterns to enable responsible companion experiences.
Type
Publication
ACL 2026 TrustNLP Workshop

Abstract

Companion systems for everyday users must balance helpfulness, trust, and social context. We introduce MASCOT, a multi-agent socio-collaborative companion framework that coordinates specialized agents to support socially grounded, trustworthy interaction.

LLM Agents Multi-Agent Systems Trustworthy AI Human-AI Interaction
Yiqiao Jin
Authors
Yiqiao Jin
Ph.D. Candidate in Computer Science
My research focuses on adaptive and efficient AI systems, with emphasis on LLM agents, agent memory, self-distillation, multimodal LLMs, and structured multi-agent intelligence.