Selective Memory for Artificial Intelligence: Write-Time Gating with Hierarchical Archiving

ArXi:2603.15994v1 Announce Type: new Retrieval-augmented generation s all content indiscriminately, degrading accuracy as noise accumulates. Parametric approaches compress knowledge into weights, precluding selective updates. Neither mirrors biological memory, which gates encoding based on salience and archives rather than deletes superseded information. We