Neuromem: A Granular Decomposition of the Streaming Lifecycle in External Memory for LLMs
This work addresses a practical gap in memory module evaluation for AI systems handling streaming data, though it is incremental as it focuses on benchmarking rather than introducing new methods.
The paper tackles the problem of evaluating External Memory Modules for LLMs under realistic streaming conditions, where memory evolves during query serving, and presents Neuromem, a testbed that benchmarks these modules across five lifecycle dimensions, reporting performance degradation as memory grows and identifying memory data structure as key to quality.
Most evaluations of External Memory Module assume a static setting: memory is built offline and queried at a fixed state. In practice, memory is streaming: new facts arrive continuously, insertions interleave with retrievals, and the memory state evolves while the model is serving queries. In this regime, accuracy and cost are governed by the full memory lifecycle, which encompasses the ingestion, maintenance, retrieval, and integration of information into generation. We present Neuromem, a scalable testbed that benchmarks External Memory Modules under an interleaved insertion-and-retrieval protocol and decomposes its lifecycle into five dimensions including memory data structure, normalization strategy, consolidation policy, query formulation strategy, and context integration mechanism. Using three representative datasets LOCOMO, LONGMEMEVAL, and MEMORYAGENTBENCH, Neuromem evaluates interchangeable variants within a shared serving stack, reporting token-level F1 and insertion/retrieval latency. Overall, we observe that performance typically degrades as memory grows across rounds, and time-related queries remain the most challenging category. The memory data structure largely determines the attainable quality frontier, while aggressive compression and generative integration mechanisms mostly shift cost between insertion and retrieval with limited accuracy gain.