🤖 AI Summary
This work addresses the limitations of traditional retrieval-augmented generation (RAG), which treats memory as a static, stateless, read-only lookup table and thus struggles to support knowledge evolution and state updates in long-term interactions. To overcome this, the authors propose the Continuum Memory Architecture (CMA), the first formalized memory system that enables stateful evolution through persistent storage, selective retention, temporal chaining, associative routing, and integration of high-level abstractions. CMA facilitates dynamic memory accumulation, updating, and disambiguation over time. Experimental results demonstrate that CMA significantly outperforms conventional RAG on tasks involving knowledge updating, temporal reasoning, associative recall, and contextual disambiguation. This advancement equips long-horizon large language model agents with essential memory capabilities, while also uncovering new challenges related to latency, semantic drift, and interpretability.
📝 Abstract
Retrieval-augmented generation (RAG) has become the default strategy for providing large language model (LLM) agents with contextual knowledge. Yet RAG treats memory as a stateless lookup table: information persists indefinitely, retrieval is read-only, and temporal continuity is absent. We define the \textit{Continuum Memory Architecture} (CMA), a class of systems that maintain and update internal state across interactions through persistent storage, selective retention, associative routing, temporal chaining, and consolidation into higher-order abstractions. Rather than disclosing implementation specifics, we specify the architectural requirements CMA imposes and show consistent behavioral advantages on tasks that expose RAG's structural inability to accumulate, mutate, or disambiguate memory. The empirical probes (knowledge updates, temporal association, associative recall, contextual disambiguation) demonstrate that CMA is a necessary architectural primitive for long-horizon agents while highlighting open challenges around latency, drift, and interpretability.