AI RESEARCH
Marco DeepResearch: Unlocking Efficient Deep Research Agents via Verification-Centric Design
arXiv CS.AI
•
ArXi:2603.28376v1 Announce Type: cross Deep research agents autonomously conduct open-ended investigations, integrating complex information retrieval with multi-step reasoning across diverse sources to solve real-world problems. To sustain this capability on long-horizon tasks, reliable verification is critical during both