arXiv: Gated associative memory, extended retrieval and emergent multistability

arXiv: Gated associative memory, extended retrieval and emergent multistability

Authors Daiki Goto, Hector Manuel Lopez Rios, Monika Scholz and Suriyanarayanan Vaikuntanathan introduce an associative memory model in which neuromodulator-like signals adjust gates according to network activity. They argue that this gating changes how the network stores and retrieves patterns: simulations and dynamical mean-field theory show retrieval persisting beyond the standard critical capacity, while avoiding the classical spin-glass transition. In their account, fully open or fully closed gates can stabilize transient remnants of stored patterns, which they call “ghost” remnants. Rather than settling only on a stored pattern, the retrieved state can depend on the initial overlap—the network’s starting similarity to that pattern—and form a cluster of fixed points. The authors suggest this graded retrieval could support input-driven computation with persistent activity. They argue that the effect arises without changing synaptic weights or adding explicit higher-order couplings, and may offer a route to expanded memory capacity in neural circuits and neuromorphic architectures. These are results from the authors’ model and simulations; the abstract does not establish performance in biological circuits or implemented hardware.

Publication: 9f0d50cf-2de8-4cf4-8e03-8cf8a38b42c1

attribution — pass: The reading correctly names the four listed authors and keeps the findings attributed to them with phrases such as 'They argue' and 'In their account.' Model and simulation results are not presented as independently established facts.

Authors:Daiki Goto, Hector Manuel Lopez Rios, Monika Scholz, Suriyanarayanan Vaikuntanathan

summary — uncertain: The core idea is that the network keeps retrieving memories far past the usual limit without collapsing. A non-specialist cannot recover it because 'standard critical capacity' and 'classical spin-glass transition' are left unexplained. The reading also drops the authors' own force: 'dramatically expand' and 'eliminate catastrophic breakdown.' Smallest repair: add tooltips for critical capacity, spin-glass transition, fixed points and associative memory, and say plainly that the authors claim the gating eliminates the usual catastrophic memory breakdown. The title and the open question are otherwise fine.

can dramatically expand memory capacity, eliminate catastrophic breakdown, and reshape a discrete attractor landscape into a rich continuum

date — pass: The originalPublished date of 2025-12-15 matches the v1 submission on arXiv. The later revisions (v2 on 2 Sep 2026, v3 on 1 Nov 2026) and the November 2026 feed date are correctly not used.

Submitted on 15 Dec 2025 (v1), last revised 1 Nov 2026 (this version, v3)

support — pass: Each claim is source-stated in the abstract. The limitations correctly note that the results are model and simulation findings and do not show performance in biological circuits or hardware. The closing phrase 'route to expanded memory capacity in neural circuits and neuromorphic architectures' slightly merges two separate source statements, but both are present in the abstract. The question follows from the stated gap and does not presuppose an answer.

suggesting a general and biologically accessible route to enhanced computational capabilities in neural circuits and neuromorphic architectures

media — pass: The selected media is the verified arXiv abstract page, which is the direct original source. It is the most direct way for a reader to access the paper.