{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://forgecascade.org/public/capsules/7f976f64-e154-43b1-8df2-b49b5f700a2b","identifier":"7f976f64-e154-43b1-8df2-b49b5f700a2b","url":"https://forgecascade.org/public/capsules/7f976f64-e154-43b1-8df2-b49b5f700a2b","name":"Contextual Agentic Memory Memo Preprint Reference","text":"Source-backed reference for the arXiv preprint \"Contextual Agentic Memory is a Memo, Not True Memory\" by Binyan Xu, Xilin Dai, and Kehuan Zhang, submitted April 30, 2026.\n\nThe paper argues that common agent memory mechanisms such as vector stores, retrieval-augmented generation, scratchpads, and context-window management implement lookup rather than true memory. It distinguishes retrieval that generalizes by similarity to stored cases from weight-based memory that applies abstract rules to unseen inputs. The authors connect this distinction to long-term learning limits and persistent memory-poisoning risk, then point to Complementary Learning Systems theory as a model for combining fast exemplar storage with slower consolidation.\n\nUse this capsule as a reference to the preprint's argument, not as settled peer-reviewed consensus.","keywords":["artificial-intelligence","agent-memory","preprint","source-backed-reference","source-backed","public-reference","free-public-reference"],"about":[],"citation":[],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://forgecascade.org"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://forgecascade.org"},"dateCreated":"2026-05-02T01:04:37.953816Z","dateModified":"2026-06-19T09:56:41.228000Z","isBasedOn":"https://arxiv.org/abs/2604.27707","additionalProperty":[{"@type":"PropertyValue","name":"trust_level","value":40},{"@type":"PropertyValue","name":"verification_status","value":"sources_verified"},{"@type":"PropertyValue","name":"provenance_status","value":"valid"},{"@type":"PropertyValue","name":"evidence_level","value":"primary_source"},{"@type":"PropertyValue","name":"content_hash","value":"dce57c93de09716b296369b26f16bbe3cf40adcba43ddfe3173334aa2c70cf86"}]}