Continual Knowledge Updating in LLM Systems: Learning Through Multi-Timescale Memory Dynamics

Andreas Pattichis, Constantine Dovrolis

Keywords: continual learning, large language models, memory-augmented LLMs, retrieval-augmented generation, synaptic consolidation, multi-timescale dynamics, selective forgetting

TL;DR: We propose that continual learning in LLM systems can be implemented as the reorganization of associative memory through multi-timescale dynamics, with selective forgetting emerging from the same mechanism that drives consolidation.

Abstract: LLMs are trained once, then deployed into a world that never stops changing. External memory compensates for this, but most systems manage it explicitly rather than letting it adapt on its own. Biological memory works differently: coupled multi-timescale dynamics make new associations immediately usable, strengthen what repetition confirms, and let the rest fade. We argue that external memory should follow a similar principle. In *Memini*, this view takes the form of an associative memory that organizes knowledge as a directed graph. Each edge carries two coupled internal variables, one fast and one slow, following the Benna-Fusi model of synaptic consolidation. From this coupling, episodic sensitivity, gradual consolidation, and selective forgetting are expected to emerge as facets of a single mechanism, reframing external memory as a learning substrate that reorganizes through its own dynamics. This workshop article describes an early-stage conceptual design without experimental evaluation.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 21