APEX-Searcher: Augmenting LLMs' Search Capabilities through Agentic Planning and Execution

Researchers introduce APEX-Searcher, a new framework that enhances large language models' search capabilities through a two-stage approach combining reinforcement learning for strategic planning and supervised fine-tuning for execution. The system addresses limitations in multi-hop question answering by decoupling retrieval processes into planning and execution phases, showing significant improvements across multiple benchmarks.

• →APEX-Searcher introduces a novel two-stage framework that separates retrieval planning from execution to improve LLM search performance.
• →The system uses reinforcement learning with decomposition-specific rewards to optimize strategic planning for complex queries.
• →Supervised fine-tuning on high-quality multi-hop trajectories enhances the model's iterative sub-task execution capabilities.
• →The framework addresses key challenges in existing RAG systems including ambiguous retrieval paths and sparse rewards in training.
• →Experimental results show significant improvements in both multi-hop RAG and task planning performance across multiple benchmarks.