cs.AINov 9, 2026

TANSAR: Editable and Attributable Correlation Matrix Memories under Correlated Keys, with Rule-Constrained Multi-Hop Knowledge Editing

Authors: Abdulhady Abas Abdulla, Jiyan Salim Mahmud, Fatemeh Daneshfar, Hussein Mohammed Ali, Mourad Oussalah, Muhammad Bilal

Organizations: University of Kurdistan Hewlêr, Erbil, Kurdistan Region, Iraq · Faculty of Informatics, Eötvös Loránd University, Budapest, Hungary · Department of Computer Engineering, University of Kurdistan, Sanandaj, Iran · Cybersecurity Department, Tishk International University, Erbil, Kurdistan Region, Iraq · 

Abstract

Correlation Matrix Memories (CMMs) store associations as sums of rank-one outer products, providing closed-form addressability, exact deletion, and linear attribution. While classical analyses assume random, orthogonal keys, semantic encoders produce correlated representations. We analyse CMMs under correlated keys using frame-theoretic coherence: (i) capacity bounds yielding an encoder-independent limit n≤d(1+ε2)n \le d(1+\varepsilon^2) for mean squared crosstalk ε2\varepsilon^2; (ii) an exact characterisation of the collateral effect of deleting an association, guaranteeing that deletion never invalidates certified associations; and (iii) a geometric transparency bound demonstrating why attribution requires candidate sets of centred keys. On this basis, we introduce Topology-Aware Neuro-Symbolic Associative Retrieval (TANSAR). TANSAR traverses a directed acyclic association graph, prunes invalid paths via sound ground Horn constraints, and injects a representation derived strictly from the verified path into the residual stream of a frozen language model, guarded by a Clopper--Pearson selective-risk gate. Controlled experiments confirm all theoretical bounds. In multi-hop knowledge editing, Horn pruning reduces answer contradictions from 25.1% to 0.5%, whereas Hopfield retrieval incurs 24% contradictions when stale associations remain stored (and 0% if stale associations are removed). An end-to-end edit transaction requires 13.6 ms on a standard CPU and 2.37 ms on GPU. The symbolic component matches gold labels across all 600 ProofWriter theories and achieves 94.0% agreement on FOLIO, demonstrating sound and efficient multi-hop associative retrieval with exact unlearning.

Figures & tables

Explore similar work

CardsList
  1. Sharp Capacity Thresholds in Linear Associative Memory: From Winner-Take-All to Listwise Retrieval

    May 6, 2026Nicholas Barnfield, Juno Kim, Eshaan Nichani +2Information RetrievalAssociative Memory

  2. Echo: KV-Cache-Free Associative Recall with Spectral Koopman Operators

    May 7, 2026Anupama Sridhar, Alexander JohansenState Space Language ModelsKoopman Operator Learning

  3. A Geometry-Based Capacity Theory for Finite-Feature Associative Memory

    Oct 6, 2026Jianhai Zhang, Donghao Zhang, Pattarawut Charatpangoon +4Associative MemoryInformation Retrieval