MiuLab

A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories

Rethinking Role-Playing Evaluation: Anonymous Benchmarking and A Systematic Study of Personality Effects

Survey Paper for Multi-Agent Interaction and Collaboration

[ICASSP 26] When Voice Matters : A study of audio llm behavior in clinical decision making

LLM for Recommender Systems

The Context Trap: Why End-to-End Audio Language Models Fail Multi-turn Dialogues

FlowSwitch: A State-Aware Framework for Workflow Transitions in Adaptive Dialogue Agents

Automating Personalized Interaction Planning for Conversational Agents

Company-Specific Knowledge Matters: Retrieval-Augmented Generation for Earnings Call Answer Rehearsal