{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,6,12]],"date-time":"2025-06-12T08:40:08Z","timestamp":1749717608691,"version":"3.41.0"},"reference-count":50,"publisher":"IOP Publishing","issue":"2","license":[{"start":{"date-parts":[[2025,6,12]],"date-time":"2025-06-12T00:00:00Z","timestamp":1749686400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,6,12]],"date-time":"2025-06-12T00:00:00Z","timestamp":1749686400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"name":"Pierre Andurand"},{"DOI":"10.13039\/501100000271","name":"Science and Technology Facilities Council","doi-asserted-by":"crossref","award":["ST\/T000686\/1"],"award-info":[{"award-number":["ST\/T000686\/1"]}],"id":[{"id":"10.13039\/501100000271","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2025,6,30]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>In this paper we present a novel approach to interpretable AI inspired by quantum field theory which we call the <jats:italic>NCoder<\/jats:italic>. The NCoder is a modified autoencoder neural network whose latent layer is prescribed to be a subset of <jats:italic>n<\/jats:italic>-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 <jats:italic>n<\/jats:italic>-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 <jats:italic>n<\/jats:italic>-point functions of the image distribution.<\/jats:p>","DOI":"10.1088\/2632-2153\/ade04c","type":"journal-article","created":{"date-parts":[[2025,6,3]],"date-time":"2025-06-03T22:58:42Z","timestamp":1748991522000},"page":"025059","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["NCoder\u2014a quantum field theory approach to encoding data"],"prefix":"10.1088","volume":"6","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-5382-1668","authenticated-orcid":true,"given":"D S","family":"Berman","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-7666-1185","authenticated-orcid":true,"given":"M S","family":"Klinger","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-6784-7779","authenticated-orcid":true,"given":"A G","family":"Stapleton","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"266","published-online":{"date-parts":[[2025,6,12]]},"reference":[{"article-title":"On the dynamics of inference and learning","year":"2022","author":"Berman","key":"mlstade04cbib1"},{"article-title":"The inverse of exact renormalization group flows as statistical inference","year":"2022","author":"Berman","key":"mlstade04cbib2"},{"key":"mlstade04cbib3","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ad0102","article-title":"Bayesian renormalization","volume":"4","author":"Berman","year":"2023","journal-title":"Mach. Learn. Sci. Tech."},{"volume":"vol 194","year":"2016","author":"Amari","key":"mlstade04cbib4"},{"key":"mlstade04cbib5","doi-asserted-by":"publisher","first-page":"1100","DOI":"10.3390\/e22101100","article-title":"An elementary introduction to information geometry","volume":"22","author":"Nielsen","year":"2020","journal-title":"Entropy"},{"article-title":"Inferring effective couplings with restricted Boltzmann machines","year":"2023","author":"Decelle","key":"mlstade04cbib6"},{"key":"mlstade04cbib7","doi-asserted-by":"publisher","first-page":"JHEP12(2023)146","DOI":"10.1007\/JHEP12(2023)146","article-title":"JAXVacua \u2013 a framework for sampling string vacua","author":"Dubey","year":"2023","journal-title":"J. High Energy Phys."},{"key":"mlstade04cbib8","first-page":"p 11","article-title":"Probing the structure of string theory vacua with genetic algorithms and reinforcement learning","author":"Cole","year":"2021"},{"article-title":"Renormalization in the neural network-quantum field theory correspondence","year":"2022","author":"Erbin","key":"mlstade04cbib9"},{"article-title":"Metric flows with neural networks","year":"2023","author":"Halverson","key":"mlstade04cbib10"},{"key":"mlstade04cbib11","first-page":"p 8","article-title":"earning from topology: cosmological parameter estimation from the large-scale structure","author":"Yip","year":"2023"},{"article-title":"Lectures on numerical and machine learning methods for approximating Ricci-flat Calabi-Yau metrics","year":"2023","author":"Anderson","key":"mlstade04cbib12"},{"article-title":"New Calabi-Yau manifolds from genetic algorithms","year":"2023","author":"Berglund","key":"mlstade04cbib13"},{"key":"mlstade04cbib14","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.108.105027","article-title":"Bootstrability in line-defect CFTs with improved truncation methods","volume":"108","author":"Niarchos","year":"2023","journal-title":"Phys. Rev. D"},{"key":"mlstade04cbib15","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.107.025005","article-title":"6D (2,0) bootstrap with the soft-actor-critic algorithm","volume":"107","author":"K\u00e1ntor","year":"2023","journal-title":"Phys. Rev. D"},{"key":"mlstade04cbib16","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.108.025003","article-title":"Renormalization group flow as optimal transport","volume":"108","author":"Cotler","year":"2023","journal-title":"Phys. Rev. D"},{"article-title":"Renormalizing diffusion models","year":"2023","author":"Cotler","key":"mlstade04cbib17"},{"article-title":"Inverse renormalization group of disordered systems","year":"2023","author":"Bachtis","key":"mlstade04cbib18"},{"article-title":"Interpreting machine learning functions as physical observables","year":"2021","author":"Aarts","key":"mlstade04cbib19"},{"article-title":"Improving energy conserving descent for machine learning: theory and practice","year":"2023","author":"Luca","key":"mlstade04cbib20"},{"key":"mlstade04cbib21","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevResearch.4.043143","article-title":"Decomposing neural networks as mappings of correlation functions","volume":"4","author":"Fischer","year":"2022","journal-title":"Phys. Rev. Res."},{"key":"mlstade04cbib22","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/abeca3","article-title":"Neural networks and quantum field theory","volume":"2","author":"Halverson","year":"2021","journal-title":"Mach. Learn. Sci. Tech."},{"article-title":"Symmetry-via-duality: invariant neural network densities from parameter-space correlators","year":"2021","author":"Maiti","key":"mlstade04cbib23"},{"article-title":"Building quantum field theories out of neurons","year":"2021","author":"Halverson","key":"mlstade04cbib24"},{"key":"mlstade04cbib25","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ad17d3","article-title":"Neural network field theories: non-Gaussianity, actions and locality","volume":"5","author":"Demirtas","year":"2024","journal-title":"Mach. Learn. Sci. Tech."},{"key":"mlstade04cbib26","doi-asserted-by":"publisher","first-page":"141","DOI":"10.1109\/MSP.2012.2211477","article-title":"The MNIST database of handwritten digit images for machine learning research","volume":"29","author":"Deng","year":"2012","journal-title":"IEEE Signal Process. Mag."},{"article-title":"Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms","year":"2017","author":"Xiao","key":"mlstade04cbib27"},{"year":"2018","author":"McCullagh","key":"mlstade04cbib28"},{"volume":"vol 26","year":"2004","author":"Wasserman","key":"mlstade04cbib29"},{"year":"2012","author":"Reed","key":"mlstade04cbib30"},{"volume":"vol 9","year":"2017","author":"Schm\u00fcdgen","key":"mlstade04cbib31"},{"key":"mlstade04cbib32","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1186\/s40488-017-0059-2","article-title":"Recent developments on the moment problem","volume":"4","author":"Lin","year":"2017","journal-title":"J. Stat. Distrib. Appl."},{"article-title":"Ten lectures on the moment problem","year":"2020","author":"Schm\u00fcdgen","key":"mlstade04cbib33"},{"year":"2013","author":"Hall","key":"mlstade04cbib34"},{"key":"mlstade04cbib35","first-page":"569","article-title":"Inverting an edgeworth expansion","author":"Hall","year":"1983","journal-title":"Ann. Stat."},{"year":"2018","author":"Peskin","key":"mlstade04cbib36"},{"article-title":"Autoencoders, unsupervised learning, and deep architectures","year":"2012","author":"Baldi","key":"mlstade04cbib37"},{"key":"mlstade04cbib38","doi-asserted-by":"publisher","first-page":"85","DOI":"10.1016\/j.neunet.2014.09.003","article-title":"Deep learning in neural networks: an overview","volume":"61","author":"Schmidhuber","year":"2015","journal-title":"Neural Netw."},{"article-title":"Tutorial on variational autoencoders","year":"2016","author":"Doersch","key":"mlstade04cbib39"},{"article-title":"The information bottleneck method","year":"2000","author":"Tishby","key":"mlstade04cbib40"},{"key":"mlstade04cbib41","first-page":"pp 599","article-title":"A practical guide to training restricted Boltzmann machines","author":"Hinton","year":"2012"},{"key":"mlstade04cbib42","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.103.036020","article-title":"Mapping machine-learned physics into a human-readable space","volume":"103","author":"Faucett","year":"2021","journal-title":"Phys. Rev. D"},{"key":"mlstade04cbib43","doi-asserted-by":"publisher","first-page":"JHEP04(2018)013","DOI":"10.1007\/JHEP04(2018)013","article-title":"Energy flow polynomials: a complete linear basis for jet substructure","author":"Komiske","year":"2018","journal-title":"J. High Energy Phys."},{"key":"mlstade04cbib44","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.103.116028","article-title":"Learning to identify electrons","volume":"103","author":"Collado","year":"2021","journal-title":"Phys. Rev. D"},{"key":"mlstade04cbib45","doi-asserted-by":"publisher","first-page":"389","DOI":"10.1016\/j.copbio.2008.06.008","article-title":"Sloppiness, robustness and evolvability in systems biology","volume":"19","author":"Daniels","year":"2008","journal-title":"Curr. Opin. Biotechnol."},{"key":"mlstade04cbib46","doi-asserted-by":"publisher","first-page":"604","DOI":"10.1126\/science.1238723","article-title":"Parameter space compression underlies emergent theories and predictive models","volume":"342","author":"Machta","year":"2013","journal-title":"Science"},{"key":"mlstade04cbib47","doi-asserted-by":"publisher","DOI":"10.1063\/1.4923066","article-title":"Perspective: Sloppiness and emergent theories in physics, biology and beyond","volume":"143","author":"Transtrum","year":"2015","journal-title":"J. Chem. Phys."},{"key":"mlstade04cbib48","doi-asserted-by":"publisher","first-page":"434","DOI":"10.3390\/e25030434","article-title":"Far from asymptopia: unbiased high-dimensional inference cannot assume unlimited data","volume":"25","author":"Abbott","year":"2023","journal-title":"Entropy"},{"article-title":"Introduction to quantum field theory Michaelmas term","year":"2010","author":"Cardy","key":"mlstade04cbib49"},{"article-title":"Information geometry of Euclidean quantum fields","year":"2023","author":"Floerchinger","key":"mlstade04cbib50"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,12]],"date-time":"2025-06-12T08:01:19Z","timestamp":1749715279000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ade04c"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,12]]},"references-count":50,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2025,6,12]]},"published-print":{"date-parts":[[2025,6,30]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/ade04c","relation":{},"ISSN":["2632-2153"],"issn-type":[{"type":"electronic","value":"2632-2153"}],"subject":[],"published":{"date-parts":[[2025,6,12]]},"assertion":[{"value":"NCoder\u2014a quantum field theory approach to encoding data","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2025 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2024-07-23","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-06-03","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-06-12","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}