S-EMBER
S-EMBER: A Large-Scale Benchmark for Streaming Egocentric Memory Retrieval
Xiaodong Wang, Xuanyi Zhao, Pedro Rodriguez, Devendra Singh Sachan, Barlas Oguz, Seungwhan Moon, et al. · 2026-07-02
- Modality
- Video & streaming
- Data
- Real data
- Task · ground truth
- QA/Ret. · Ans.+Ev.+Time
- Scope
- Memory-specific benchmark
What it tests
- CECross-modal evidencePartially evaluated
- LELong-range evidenceDirectly evaluated
- SCState changesPartially evaluated
- CFConflicting memoriesNot evaluated
- UQUnanswerable queriesNot evaluated
- MAMemory-grounded actionNot evaluated
- TLTemporal localizationDirectly evaluated
- TRTemporal relationsDirectly evaluated
- MEMulti-event integrationDirectly evaluated
Survey notes
Causal look-back questions require past visual evidence that is unavailable in the current frame, with supporting temporal intervals.