ARGORA: Orchestrated Argumentation for Causally Grounded LLM Reasoning and Decision Making

2026-01-29

AI Summary

This work addresses a key limitation in existing multi-expert large language model systems, which, despite integrating diverse perspectives, rely on simplistic aggregation mechanisms that obscure the influence of individual arguments on final decisions. To overcome this, the authors propose an explicit argumentation graph framework that structures support and attack relations among experts and, for the first time, formalizes them as a causal graph. This enables counterfactual interventions on specific arguments and facilitates decision attribution through causal reasoning. Additionally, a consistency alignment mechanism is introduced to automatically refine internal reasoning using external feedback. Experimental results demonstrate that the approach achieves competitive accuracy across multiple benchmarks and open-ended scenarios, particularly excelling in correcting initial expert disagreements while providing interpretable causal diagnoses.

Technology Category

• Knowledge Representation and Reasoning: Argumentation
• Multiagent Systems: Agreement, Argumentation & Negotiation
• Reasoning under Uncertainty: Causality

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Existing multi-expert LLM systems gather diverse perspectives but combine them through simple aggregation, obscuring which arguments drove the final decision. We introduce ARGORA, a framework that organizes multi-expert discussions into explicit argumentation graphs showing which arguments support or attack each other. By casting these graphs as causal models, ARGORA can systematically remove individual arguments and recompute outcomes, identifying which reasoning chains were necessary and whether decisions would change under targeted modifications. We further introduce a correction mechanism that aligns internal reasoning with external judgments when they disagree. Across diverse benchmarks and an open-ended use case, ARGORA achieves competitive accuracy and demonstrates corrective behavior: when experts initially disagree, the framework resolves disputes toward correct answers more often than it introduces new errors, while providing causal diagnostics of decisive arguments.

Problem

• multi-expert LLM systems
• argumentation
• causal reasoning
• decision justification
• reasoning transparency

Innovation

• argumentation graphs
• causal reasoning
• multi-expert LLM
• decision correction
• reasoning diagnostics