mangopy/Deep-Research-Survey

Organizes research papers on Deep Research systems across four key architectural components—query planning, information acquisition, agentic reinforcement learning, and supervised fine-tuning—with structured reading lists and benchmark resources. Provides a systematic taxonomy covering knowledge boundary detection, retrieval timing optimization, and end-to-end training paradigms for LLM-based research agents. Includes a peer-reviewed survey (arXiv:2512.02038) with implementation roadmaps, practical techniques, and emerging challenges in autonomous deep research systems.

308 stars.