Dual Hierarchical Dialogue Policy Learning for Legal Inquisitive Conversational Agents

2026-05-13

AI Summary

This work addresses the limitation of conventional dialogue systems, which typically respond passively and struggle to proactively elicit critical information in high-stakes scenarios. The paper introduces, for the first time, the concept of an "interrogative dialogue agent" tailored to U.S. Supreme Court oral arguments. It proposes a hierarchical reinforcement learning framework comprising a high-level policy that plans interrogative goals and a low-level policy that generates specific utterances, enabling coordinated, goal-directed information seeking. Evaluated on a real-world Supreme Court dataset, the approach significantly outperforms multiple baselines across several metrics, demonstrating superior performance in both information acquisition capability and strategic dialogue effectiveness.

Abstract

Most existing dialogue systems are user-driven, primarily designed to fulfill user requests. However, in many critical real-world scenarios, a conversational agent must proactively extract information to achieve its own objectives rather than merely respond. To address this gap, we introduce \emph{Inquisitive Conversational Agents (ICAs)} and develop an ICA specifically tailored to U.S. Supreme Court oral arguments. We propose a Dual Hierarchical Reinforcement Learning framework featuring two cooperating RL agents, each with its own policy, to coordinate strategic dialogue management and fine-grained utterance generation. By learning when and how to ask probing questions, the agent emulates judicial questioning patterns and systematically uncovers crucial information to fulfill its legal objectives. Evaluations on a U.S. Supreme Court dataset show that our method outperforms various baselines across multiple metrics. It represents an important first step toward broader high-stakes, domain-specific applications.

Problem

• Inquisitive Conversational Agents
• proactive information extraction
• dialogue policy
• judicial questioning
• goal-oriented dialogue

Innovation

• Inquisitive Conversational Agents
• Dual Hierarchical Reinforcement Learning
• Proactive Questioning
• Legal Dialogue Systems
• Judicial Oral Arguments