1 paper · 1 filter
Ekko He, Jack Hong, Shilin Yan +8
Recent multimodal deep search agents have advanced tool use and visual content retrieval, yet they still struggle to convert intermediate visual evidence into actionable guidance f…