Reasoning About Intent for Ambiguous Requests

2025-11-13

AI Summary

Large language models (LLMs) often implicitly commit to a single interpretation of ambiguous user queries, leading to intent misidentification and safety risks. Method: We propose a structured multi-intent response framework that generates, in a single forward pass, multiple semantically plausible and mutually exclusive interpretations of the user’s intent—each paired with a corresponding answer. Our approach employs a reinforcement learning paradigm with a novel multi-answer consistency reward function, leveraging multiple valid outputs as weak supervision signals. Contribution/Results: Unlike iterative clarification or single-answer generation, our method achieves high response efficiency, transparency, and coverage without requiring additional user interaction. Experiments on dialogue QA and semantic parsing tasks demonstrate a +18.7% improvement in effective answer coverage. Human evaluation confirms 92.4% consistency between inferred intents and their associated answers, validating both efficacy and practical utility.

Technology Category

• Natural Language Processing: Question Answering
• Machine Learning: Large Multimodal Models (LMMs)
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Large language models often respond to ambiguous requests by implicitly committing to one interpretation. Intent misunderstandings can frustrate users and create safety risks. To address this, we propose generating multiple interpretation-answer pairs in a single structured response to ambiguous requests. Our models are trained with reinforcement learning and customized reward functions using multiple valid answers as supervision. Experiments on conversational question answering and semantic parsing demonstrate that our method achieves higher coverage of valid answers than baseline approaches. Human evaluation confirms that predicted interpretations are highly aligned with their answers. Our approach promotes transparency with explicit interpretations, achieves efficiency by requiring only one generation step, and supports downstream applications through its structured output format.

Problem

• Addressing ambiguous requests that cause intent misunderstandings in language models
• Improving coverage of valid interpretations for ambiguous user queries
• Enhancing transparency and safety through explicit interpretation-answer pairs

Innovation

• Generates multiple interpretation-answer pairs in one response
• Trains models with reinforcement learning and custom rewards
• Uses structured output for transparency and efficiency