Perspectra: Choosing Your Experts Enhances Critical Thinking in Multi-Agent Research Ideation
2025/09/24 by Liu, Yiren, Shah, Viraj, Suh, Sangho +3 · 1 citation
#Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #FOS: Computer and information sciences #Human-Computer Interaction (cs.HC)
paper · doi:10.48550/arxiv.2509.20553
Abstract
Recent advances in multi-agent systems (MAS) enable tools for information search and ideation by assigning personas to agents. However, how users can effectively control, steer, and critically evaluate collaboration among multiple domain-expert agents remains underexplored. We present Perspectra, an interactive MAS that visualizes and structures deliberation among LLM agents via a forum-style interface, supporting @-mention to invite targeted agents, threading for parallel exploration, with a real-time mind map for visualizing arguments and rationales. In a within-subjects study with 18 participants, we compared Perspectra to a group-chat baseline as they developed research proposals. Our findings show that Perspectra significantly increased the frequency and depth of critical-thinking behaviors, elicited more interdisciplinary replies, and led to more frequent proposal revisions than the group chat condition. We discuss implications for designing multi-agent tools that scaffold critical thinking by supporting user control over multi-agent adversarial discourse.
Citations
- Exploring Design of Multi-Agent LLM Dialogues for Research Ideation
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Scaffolding Recursive Divergence and Convergence in Story Ideation
- From Conversation to Orchestration: HCI Challenges and Opportunities in Interactive Multi-Agentic Systems
- Deep Research Agents: A Systematic Examination And Roadmap
- Needling Through the Threads: A Visualization Tool for Navigating Threaded Online Discussions
- Large Language Model-Empowered Interactive Load Forecasting
- IRIS: Interactive Research Ideation System for Accelerating Scientific Discovery
- DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environments
- From Consumption to Collaboration: Measuring Interaction Patterns to Augment Human Cognition in Open-Ended Tasks
- MDTeamGPT: A Self-Evolving LLM-based Multi-Agent Framework for Multi-Disciplinary Team Medical Consultation
- DBox: Scaffolding Algorithmic Programming Learning through Learner-LLM Co-Decomposition
- Proactive Conversational Agents with Inner Thoughts
- Argumentative Experience: Reducing Confirmation Bias on Controversial Issues through LLM-Generated Multi-Persona Debates
- Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks
- Navigating the Unknown: A Chat-Based Collaborative Interface for Personalized Exploratory Tasks
- Self-Preference Bias in LLM-as-a-Judge
- Persona-L has Entered the Chat: Leveraging LLM and Ability-based Framework for Personas of People with Complex Needs
- PersonaFlow: Designing LLM-Simulated Expert Perspectives for Enhanced Research Ideation
- Into the Unknown Unknowns: Engaged Human Learning through Participation in Language Model Agent Conversations
- Scaling Synthetic Data Creation with 1,000,000,000 Personas
- Simulating Classroom Education with LLM-Empowered Agents
- Conversational Agents as Catalysts for Critical Thinking: Challenging Design Fixation in Group Design
- Enabling Generative Design Tools with LLM Agents for Mechanical Computation Devices: A Case Study
- LLM Evaluators Recognize and Favor Their Own Generations
- Shaping Human-AI Collaboration: Varied Scaffolding Levels in Co-writing with Language Models
- Form-From: A Design Space of Social Media Systems
- SocraSynth: Multi-LLM Reasoning with Conditional Statistics
- Learning to Break: Knowledge-Enhanced Reasoning in Multi-Agent Debate System
- Beyond ChatBots: ExploreLLM for Structured Thoughts and Personalized Model Responses
- AI Supported Degradation of the Self Concept: A Theoretical Framework Grounded in Established Cognitive and Computational Mechanisms
- CoQuest: Exploring Research Question Co-Creation with an LLM-based Agent
- ChoiceMates: Supporting Unfamiliar Online Decision-Making with Multi-Agent Conversational Interactions
- Fostering User Engagement in the Critical Reflection of Arguments
- CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society
- Distributing Accountability, Not Capability: Phase Separation and the LLM Workflow Quadrant in Autonomous AI Agent Architectures
- OpenAlex Snapshot
- AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts
- Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy
- Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy
- The Rapidly Changing Landscape of Conversational Agents
- Collaborative discourse, argumentation, and learning: Preface and literature review
- Trust in Automation: Designing for Appropriate Reliance
- MetaCrit: A Critical Thinking Framework for Self-Regulated LLM Reasoning
- Human-LLM Compound System for Scientific Ideation through Facet Recombination and Novelty Evaluation
- The Semantic Scholar Open Data Platform
- Productive friction: How conflict in student teaching creates opportunities for learning at the boundary
Cited by
Related