arXiv:2606.24040v1 Announce Type: new 
Abstract: MeMo proposes language models with explicit multi-layer correlation matrix memories (CMMs), where memorization, retrieval, and forgetting are architectural operations. This paper asks how such memories can reduce the need for retraining when knowledge changes. For changes expressible as MeMo memory associations, the model's accessible knowledge can be updated by editing explicit memories rather than retraining the whole model. We propose a version-aware operation layer in which high-level operations such as replace, obsolete, keep-history, rollback, and trace are compiled into MeMo-native primitive calls over sequences and tokens. The key observation is that a version-aware operation is rarely a single MeMo association. It is an ordered transaction of primitive edits, for example forgetting one sequence-token chain, memorizing another, preserving a historical chain, and recording an inverse program. The framework introduces two auxiliary CMMs: a Version CMM (V-CMM) for mapping version transitions to transaction handles, and a Transaction CMM (T-CMM) for storing reusable change contents and inverse programs. It supports both direct sequence-level edits and structured diff-level inputs, and outlines an evaluation route for update success, rollback, traceability, locality, and transaction reuse.

يقدم مقال جديد بعنوان 'نحو عمليات واعية بالإصدارات وذاكرات معاملات متعددة الطبقات لـ MeMo' إطارًا لـ MeMo، الذي يستخدم ذاكرات مصفوفة الارتباط متعددة الطبقات (CMMs) لتحسين نماذج اللغة. يسمح هذا الإطار بتحديث المعرفة من خلال تحرير الذاكرة بدلاً من إعادة التدريب الكامل، مقترحًا طبقة عمليات واعية بالإصدارات لإدارة فعالة للذاكرة.

Un nuevo artículo titulado 'Hacia operaciones y memorias de transacción conscientes de versiones para MeMo de múltiples capas' presenta un marco para MeMo, que utiliza memorias de matriz de correlación de múltiples capas (CMMs) para mejorar los modelos de lenguaje. Este marco permite la actualización del conocimiento mediante la edición de memoria en lugar de un reentrenamiento completo, proponiendo una capa de operación consciente de versiones para una gestión eficiente de la memoria.

Un nouvel article intitulé 'Vers des opérations et des mémoires de transaction conscientes des versions pour MeMo multi-couches' présente un cadre pour MeMo, qui utilise des mémoires de corrélation multi-couches explicites (CMMs) pour améliorer les modèles de langage. Ce cadre permet la mise à jour des connaissances par l'édition de mémoire plutôt que par un réentraînement complet, proposant une couche d'opération consciente des versions pour une gestion efficace de la mémoire.

A new paper titled 'Towards Version-aware Operations and Transaction Memories for Multi-layer MeMo' introduces a framework for MeMo, which utilizes explicit multi-layer correlation matrix memories (CMMs) to enhance language models. This framework allows for the updating of knowledge through memory editing rather than complete retraining, proposing a version-aware operation layer for efficient memory management.