The memory tesseract: Mathematical equivalence between composite and separate storage memory models

Matthew Kelly, D. J.K. Mewhort, Robert L. West

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Computational memory models can explain the behaviour of human memory in diverse experimental paradigms. But research has produced a profusion of competing models, and, as different models focus on different phenomena, there is no best model. However, by examining commonalities among models, we can move towards theoretical unification. Computational memory models can be grouped into composite and separate storage models. We prove that MINERVA 2, a separate storage model of long-term memory, is mathematically equivalent to composite storage memory implemented as a fourth order tensor, and approximately equivalent to a fourth-order tensor compressed into a holographic vector. Building of these demonstrations, we show that MINERVA 2 and related separate storage models can be implemented in neurons. Our work clarifies the relationship between composite and separate storage models of memory, and thereby moves memory models a step closer to theoretical unification.

Original languageEnglish (US)
Pages (from-to)142-155
Number of pages14
JournalJournal of Mathematical Psychology
Volume77
DOIs
StatePublished - Apr 1 2017

Fingerprint

Tesseract
Memory Model
Composite
Equivalence
Data storage equipment
Composite materials
Unification
Model
Computational Model
Long-Term Memory
Fourth Order
Tensor
Tensors
Memory Term
Neurons
Neuron
Research
Paradigm

All Science Journal Classification (ASJC) codes

  • Psychology(all)
  • Applied Mathematics

Cite this

@article{e9cfa9d7dd8f494390d78180a6d29231,
title = "The memory tesseract: Mathematical equivalence between composite and separate storage memory models",
abstract = "Computational memory models can explain the behaviour of human memory in diverse experimental paradigms. But research has produced a profusion of competing models, and, as different models focus on different phenomena, there is no best model. However, by examining commonalities among models, we can move towards theoretical unification. Computational memory models can be grouped into composite and separate storage models. We prove that MINERVA 2, a separate storage model of long-term memory, is mathematically equivalent to composite storage memory implemented as a fourth order tensor, and approximately equivalent to a fourth-order tensor compressed into a holographic vector. Building of these demonstrations, we show that MINERVA 2 and related separate storage models can be implemented in neurons. Our work clarifies the relationship between composite and separate storage models of memory, and thereby moves memory models a step closer to theoretical unification.",
author = "Matthew Kelly and Mewhort, {D. J.K.} and West, {Robert L.}",
year = "2017",
month = "4",
day = "1",
doi = "10.1016/j.jmp.2016.10.006",
language = "English (US)",
volume = "77",
pages = "142--155",
journal = "Journal of Mathematical Psychology",
issn = "0022-2496",
publisher = "Academic Press Inc.",

}

The memory tesseract : Mathematical equivalence between composite and separate storage memory models. / Kelly, Matthew; Mewhort, D. J.K.; West, Robert L.

In: Journal of Mathematical Psychology, Vol. 77, 01.04.2017, p. 142-155.

Research output: Contribution to journalArticle

TY - JOUR

T1 - The memory tesseract

T2 - Mathematical equivalence between composite and separate storage memory models

AU - Kelly, Matthew

AU - Mewhort, D. J.K.

AU - West, Robert L.

PY - 2017/4/1

Y1 - 2017/4/1

N2 - Computational memory models can explain the behaviour of human memory in diverse experimental paradigms. But research has produced a profusion of competing models, and, as different models focus on different phenomena, there is no best model. However, by examining commonalities among models, we can move towards theoretical unification. Computational memory models can be grouped into composite and separate storage models. We prove that MINERVA 2, a separate storage model of long-term memory, is mathematically equivalent to composite storage memory implemented as a fourth order tensor, and approximately equivalent to a fourth-order tensor compressed into a holographic vector. Building of these demonstrations, we show that MINERVA 2 and related separate storage models can be implemented in neurons. Our work clarifies the relationship between composite and separate storage models of memory, and thereby moves memory models a step closer to theoretical unification.

AB - Computational memory models can explain the behaviour of human memory in diverse experimental paradigms. But research has produced a profusion of competing models, and, as different models focus on different phenomena, there is no best model. However, by examining commonalities among models, we can move towards theoretical unification. Computational memory models can be grouped into composite and separate storage models. We prove that MINERVA 2, a separate storage model of long-term memory, is mathematically equivalent to composite storage memory implemented as a fourth order tensor, and approximately equivalent to a fourth-order tensor compressed into a holographic vector. Building of these demonstrations, we show that MINERVA 2 and related separate storage models can be implemented in neurons. Our work clarifies the relationship between composite and separate storage models of memory, and thereby moves memory models a step closer to theoretical unification.

UR - http://www.scopus.com/inward/record.url?scp=85006786172&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85006786172&partnerID=8YFLogxK

U2 - 10.1016/j.jmp.2016.10.006

DO - 10.1016/j.jmp.2016.10.006

M3 - Article

AN - SCOPUS:85006786172

VL - 77

SP - 142

EP - 155

JO - Journal of Mathematical Psychology

JF - Journal of Mathematical Psychology

SN - 0022-2496

ER -