NCoder—a quantum field theory approach to encoding data

In this paper we present a novel approach to interpretable AI inspired by quantum field theory which we call the NCoder . The NCoder is a modified autoencoder neural network whose latent layer is prescribed to be a subset of n -point correlation functions. Regarding images as draws from a lattice fi...

Full description

Saved in:
Bibliographic Details
Main Authors: D S Berman, M S Klinger, A G Stapleton
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:Machine Learning: Science and Technology
Subjects:
Online Access:https://doi.org/10.1088/2632-2153/ade04c
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850163333387255808
author D S Berman
M S Klinger
A G Stapleton
author_facet D S Berman
M S Klinger
A G Stapleton
author_sort D S Berman
collection DOAJ
description In this paper we present a novel approach to interpretable AI inspired by quantum field theory which we call the NCoder . The NCoder is a modified autoencoder neural network whose latent layer is prescribed to be a subset of n -point correlation functions. Regarding images as draws from a lattice field theory, this architecture mimics the task of perturbatively constructing the effective action of the theory order by order in an expansion using Feynman diagrams. Alternatively, the NCoder may be regarded as simulating the procedure of statistical inference whereby high dimensional data is first summarized in terms of several lower dimensional summary statistics (here the n -point correlation functions), and subsequent out-of-sample data is generated by inferring the data generating distribution from these statistics. In this way the NCoder suggests a fascinating correspondence between perturbative renormalizability and the sufficiency of models. We demonstrate the efficacy of the NCoder by applying it to the generation of MNIST images, and find that generated images can be correctly classified using only information from the first three n -point functions of the image distribution.
format Article
id doaj-art-513b293607df4f2a8ec41ec0b7c31daf
institution OA Journals
issn 2632-2153
language English
publishDate 2025-01-01
publisher IOP Publishing
record_format Article
series Machine Learning: Science and Technology
spelling doaj-art-513b293607df4f2a8ec41ec0b7c31daf2025-08-20T02:22:19ZengIOP PublishingMachine Learning: Science and Technology2632-21532025-01-016202505910.1088/2632-2153/ade04cNCoder—a quantum field theory approach to encoding dataD S Berman0https://orcid.org/0000-0002-5382-1668M S Klinger1https://orcid.org/0009-0001-7666-1185A G Stapleton2https://orcid.org/0009-0009-6784-7779Centre for Theoretical Physics, Queen Mary University of London , E1 4NS London, United KingdomDepartment of Physics, University of Illinois , Urbana, IL 61801, United States of AmericaCentre for Theoretical Physics, Queen Mary University of London , E1 4NS London, United KingdomIn this paper we present a novel approach to interpretable AI inspired by quantum field theory which we call the NCoder . The NCoder is a modified autoencoder neural network whose latent layer is prescribed to be a subset of n -point correlation functions. Regarding images as draws from a lattice field theory, this architecture mimics the task of perturbatively constructing the effective action of the theory order by order in an expansion using Feynman diagrams. Alternatively, the NCoder may be regarded as simulating the procedure of statistical inference whereby high dimensional data is first summarized in terms of several lower dimensional summary statistics (here the n -point correlation functions), and subsequent out-of-sample data is generated by inferring the data generating distribution from these statistics. In this way the NCoder suggests a fascinating correspondence between perturbative renormalizability and the sufficiency of models. We demonstrate the efficacy of the NCoder by applying it to the generation of MNIST images, and find that generated images can be correctly classified using only information from the first three n -point functions of the image distribution.https://doi.org/10.1088/2632-2153/ade04cneural networkcorrelation functionsquantum field theoryperturbation theoryedgeworthmoment problem
spellingShingle D S Berman
M S Klinger
A G Stapleton
NCoder—a quantum field theory approach to encoding data
Machine Learning: Science and Technology
neural network
correlation functions
quantum field theory
perturbation theory
edgeworth
moment problem
title NCoder—a quantum field theory approach to encoding data
title_full NCoder—a quantum field theory approach to encoding data
title_fullStr NCoder—a quantum field theory approach to encoding data
title_full_unstemmed NCoder—a quantum field theory approach to encoding data
title_short NCoder—a quantum field theory approach to encoding data
title_sort ncoder a quantum field theory approach to encoding data
topic neural network
correlation functions
quantum field theory
perturbation theory
edgeworth
moment problem
url https://doi.org/10.1088/2632-2153/ade04c
work_keys_str_mv AT dsberman ncoderaquantumfieldtheoryapproachtoencodingdata
AT msklinger ncoderaquantumfieldtheoryapproachtoencodingdata
AT agstapleton ncoderaquantumfieldtheoryapproachtoencodingdata