建立AI时代“度量衡”：OPPO联合OpenKG推出首个端侧AI记忆评测基准MobileMem
Ranking
Overall
75
Content
85
Popularity
N/A
No observed public metrics; popularity remains neutral/archived.
Merged summary
TL;DR - OPPO and OpenKG introduced MobileMem, an open benchmark for evaluating long-term memory in on-device AI agents. It aims to standardize comparisons under mobile constraints such as limited compute, storage, and privacy requirements.
- MobileMem uses histories totaling roughly one million tokens to simulate long-term, personalized device use.
- It evaluates event modeling and retrieval across time, applications, and modalities; multimodal association and reasoning; and evolving user preferences, relationships, and event states.
- The framework supports different agent architectures and tests components including task understanding, memory retrieval, search, and execution.
- The MobileMem 1.0 white paper is available, while datasets, tasks, and evaluation tools are planned for phased open-source release.
Sources (1)
建立AI时代“度量衡”：OPPO联合OpenKG推出首个端侧AI记忆评测基准MobileMem
Public signals
N/A
TL;DR - OPPO and OpenKG introduced MobileMem, an open benchmark for evaluating long-term memory in on-device AI agents. It aims to standardize comparisons under mobile constraints such as limited compute, storage, and privacy requirements.
- MobileMem uses histories totaling roughly one million tokens to simulate long-term, personalized device use.
- It evaluates event modeling and retrieval across time, applications, and modalities; multimodal association and reasoning; and evolving user preferences, relationships, and event states.
- The framework supports different agent architectures and tests components including task understanding, memory retrieval, search, and execution.
- The MobileMem 1.0 white paper is available, while datasets, tasks, and evaluation tools are planned for phased open-source release.