The VisSearch benchmark evaluates long-horizon visual search in multimodal agents.
VisSearch-Bench tests 16 multimodal agents on 204 visual search tasks, showing only 26.47% full success due to weak long-horizon visual reasoning.
VisSearch-Bench tests 16 multimodal agents on 204 visual search tasks, showing only 26.47% full success due to weak long-horizon visual reasoning.
Survey maps multimodal agent memory representation and management, separating content updates from policy learning for static, dynamic, evolving agents
ViewCraft evolves evidence interfaces and reading procedures for multimodal agents, using failures and replay to gain 6.1 points on eight benchmarks