IntPro: A Proxy Agent for Context-Aware Intent Understanding via Retrieval-conditioned Inference

Researchers propose IntPro, a new AI proxy agent that improves intent understanding by learning from individual user patterns through retrieval-conditioned inference. The system uses historical intent data and specialized training methods to better interpret user intentions in context-aware scenarios.

• →IntPro addresses limitations in current LLM intent understanding by incorporating user history and contextual reasoning.
• →The system stores intent explanations in individual libraries for retrieval-based pattern matching.
• →Training involves supervised fine-tuning and multi-turn Group Relative Policy Optimization with tool-aware rewards.
• →Experiments across three scenarios show strong performance improvements in intent understanding capabilities.
• →The approach moves beyond static recognition to dynamic, personalized intent inference.