MemEye
MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory
Minghao Guo, Qingyue Jiao, Zeru Shi, Yihao Quan, Boxuan Zhang, Danrui Li, et al. · 2026-05-14
- Modality
- Image
- Data
- Hybrid (real + synthetic/simulated)
- Task · ground truth
- QA · Ans.+Ev.
- Scope
- Memory-specific benchmark
What it tests
- CECross-modal evidenceDirectly evaluated
- LELong-range evidenceDirectly evaluated
- SCState changesDirectly evaluated
- CFConflicting memoriesNot evaluated
- UQUnanswerable queriesNot evaluated
- MAMemory-grounded actionNot evaluated
- FGFine-grained visual detailDirectly evaluated
- ETEntity trackingPartially evaluated
- VCVisual state changeDirectly evaluated
Survey notes
Mirrored questions require retained visual details at different granularities and reasoning about the valid state of evolving visual evidence.