000874632 001__ 874632 000874632 005__ 20240313103111.0 000874632 0247_ $$2doi$$a10.1088/1751-8121/ab82dd 000874632 0247_ $$2ISSN$$a0022-3689 000874632 0247_ $$2ISSN$$a0301-0015 000874632 0247_ $$2ISSN$$a0305-4470 000874632 0247_ $$2ISSN$$a1361-6447 000874632 0247_ $$2ISSN$$a1751-8113 000874632 0247_ $$2ISSN$$a1751-8121 000874632 0247_ $$2ISSN$$a2051-2155 000874632 0247_ $$2ISSN$$a2051-2163 000874632 0247_ $$2Handle$$a2128/25517 000874632 0247_ $$2WOS$$aWOS:000561625100001 000874632 0247_ $$2altmetric$$aaltmetric:89345132 000874632 037__ $$aFZJ-2020-01552 000874632 082__ $$a530 000874632 1001_ $$0P:(DE-Juel1)156459$$aDahmen, David$$b0$$eCorresponding author 000874632 245__ $$aCapacity of the covariance perceptron 000874632 260__ $$aBristol$$bIOP Publ.$$c2020 000874632 3367_ $$2DRIVER$$aarticle 000874632 3367_ $$2DataCite$$aOutput Types/Journal article 000874632 3367_ $$0PUB:(DE-HGF)16$$2PUB:(DE-HGF)$$aJournal Article$$bjournal$$mjournal$$s1599746003_22307 000874632 3367_ $$2BibTeX$$aARTICLE 000874632 3367_ $$2ORCID$$aJOURNAL_ARTICLE 000874632 3367_ $$00$$2EndNote$$aJournal Article 000874632 520__ $$aThe classical perceptron is a simple neural network that performs a binary classification by a linear mapping between static inputs and outputs and application of a threshold. For small inputs, neural networks in a stationary state also perform an effectively linear input-output transformation, but of an entire time series. Choosing the temporal mean of the time series as the feature for classification, the linear transformation of the network with subsequent thresholding is equivalent to the classical perceptron. Here we show that choosing covariances of time series as the feature for classification maps the neural network to what we call a 'covariance perceptron'; a mapping between covariances that is bilinear in terms of weights. By extending Gardner's theory of connections to this bilinear problem, using a replica symmetric mean-field theory, we compute the pattern and information capacities of the covariance perceptron in the infinite-size limit. Closed-form expressions reveal superior pattern capacity in the binary classification task compared to the classical perceptron in the case of a high-dimensional input and low-dimensional output. For less convergent networks, the mean perceptron classifies a larger number of stimuli. However, since covariances span a much larger input and output space than means, the amount of stored information in the covariance perceptron exceeds the classical counterpart. For strongly convergent connectivity it is superior by a factor equal to the number of input neurons. Theoretical calculations are validated numerically for finite size systems using a gradient-based optimization of a soft-margin, as well as numerical solvers for the NP hard quadratically constrained quadratic programming problem, to which training can be mapped. 000874632 536__ $$0G:(DE-HGF)POF3-571$$a571 - Connectivity and Activity (POF3-571)$$cPOF3-571$$fPOF III$$x0 000874632 536__ $$0G:(DE-HGF)POF3-574$$a574 - Theory, modelling and simulation (POF3-574)$$cPOF3-574$$fPOF III$$x1 000874632 536__ $$0G:(DE-Juel1)HGF-SMHB-2014-2018$$aMSNN - Theory of multi-scale neuronal networks (HGF-SMHB-2014-2018)$$cHGF-SMHB-2014-2018$$fMSNN$$x2 000874632 536__ $$0G:(EU-Grant)785907$$aHBP SGA2 - Human Brain Project Specific Grant Agreement 2 (785907)$$c785907$$fH2020-SGA-FETFLAG-HBP-2017$$x3 000874632 536__ $$0G:(DE-82)EXS-SF-neuroIC002$$aneuroIC002 - Recurrence and stochasticity for neuro-inspired computation (EXS-SF-neuroIC002)$$cEXS-SF-neuroIC002$$x4 000874632 588__ $$aDataset connected to CrossRef 000874632 7001_ $$0P:(DE-HGF)0$$aGilson, Matthieu$$b1 000874632 7001_ $$0P:(DE-Juel1)144806$$aHelias, Moritz$$b2$$ufzj 000874632 773__ $$0PERI:(DE-600)1363010-6$$a10.1088/1751-8121/ab82dd$$n35$$p354002$$tJournal of physics / A$$v53$$x0022-3689$$y2020 000874632 8564_ $$uhttps://juser.fz-juelich.de/record/874632/files/8157308.pdf 000874632 8564_ $$uhttps://juser.fz-juelich.de/record/874632/files/Dahmen_2020_J._Phys._A%20_Math._Theor._53_354002-2.pdf$$yOpenAccess 000874632 8564_ $$uhttps://juser.fz-juelich.de/record/874632/files/Dahmen_2020_J._Phys._A%20_Math._Theor._53_354002-2.pdf?subformat=pdfa$$xpdfa$$yOpenAccess 000874632 8564_ $$uhttps://juser.fz-juelich.de/record/874632/files/8157308.pdf?subformat=pdfa$$xpdfa 000874632 8767_ $$d2020-03-20$$eHybrid-OA$$jOffsetting$$lOffsetting: IOP$$pJPhysA-113033.R1 000874632 8767_ $$88157308$$92020-09-10$$d2020-09-10$$eColour charges$$jZahlung erfolgt$$pJPhysA-113033.R1 000874632 909CO $$ooai:juser.fz-juelich.de:874632$$pdnbdelivery$$popenCost$$pec_fundedresources$$pVDB$$pdriver$$pOpenAPC$$popen_access$$popenaire 000874632 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)156459$$aForschungszentrum Jülich$$b0$$kFZJ 000874632 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)144806$$aForschungszentrum Jülich$$b2$$kFZJ 000874632 9131_ $$0G:(DE-HGF)POF3-571$$1G:(DE-HGF)POF3-570$$2G:(DE-HGF)POF3-500$$3G:(DE-HGF)POF3$$4G:(DE-HGF)POF$$aDE-HGF$$bKey Technologies$$lDecoding the Human Brain$$vConnectivity and Activity$$x0 000874632 9131_ $$0G:(DE-HGF)POF3-574$$1G:(DE-HGF)POF3-570$$2G:(DE-HGF)POF3-500$$3G:(DE-HGF)POF3$$4G:(DE-HGF)POF$$aDE-HGF$$bKey Technologies$$lDecoding the Human Brain$$vTheory, modelling and simulation$$x1 000874632 9141_ $$y2020 000874632 915__ $$0StatID:(DE-HGF)0200$$2StatID$$aDBCoverage$$bSCOPUS 000874632 915__ $$0LIC:(DE-HGF)CCBY4$$2HGFVOC$$aCreative Commons Attribution CC BY 4.0 000874632 915__ $$0StatID:(DE-HGF)0600$$2StatID$$aDBCoverage$$bEbsco Academic Search 000874632 915__ $$0StatID:(DE-HGF)0100$$2StatID$$aJCR$$bJ PHYS A-MATH THEOR : 2017 000874632 915__ $$0StatID:(DE-HGF)0150$$2StatID$$aDBCoverage$$bWeb of Science Core Collection 000874632 915__ $$0StatID:(DE-HGF)0110$$2StatID$$aWoS$$bScience Citation Index 000874632 915__ $$0StatID:(DE-HGF)0111$$2StatID$$aWoS$$bScience Citation Index Expanded 000874632 915__ $$0StatID:(DE-HGF)9900$$2StatID$$aIF < 5 000874632 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess 000874632 915__ $$0StatID:(DE-HGF)0030$$2StatID$$aPeer Review$$bASC 000874632 915__ $$0StatID:(DE-HGF)1150$$2StatID$$aDBCoverage$$bCurrent Contents - Physical, Chemical and Earth Sciences 000874632 915__ $$0StatID:(DE-HGF)0430$$2StatID$$aNational-Konsortium 000874632 915__ $$0StatID:(DE-HGF)0300$$2StatID$$aDBCoverage$$bMedline 000874632 915__ $$0StatID:(DE-HGF)0420$$2StatID$$aNationallizenz 000874632 915__ $$0StatID:(DE-HGF)0199$$2StatID$$aDBCoverage$$bClarivate Analytics Master Journal List 000874632 9201_ $$0I:(DE-Juel1)INM-6-20090406$$kINM-6$$lComputational and Systems Neuroscience$$x0 000874632 9201_ $$0I:(DE-Juel1)IAS-6-20130828$$kIAS-6$$lTheoretical Neuroscience$$x1 000874632 9201_ $$0I:(DE-Juel1)INM-10-20170113$$kINM-10$$lJara-Institut Brain structure-function relationships$$x2 000874632 9801_ $$aAPC 000874632 9801_ $$aFullTexts 000874632 980__ $$ajournal 000874632 980__ $$aVDB 000874632 980__ $$aI:(DE-Juel1)INM-6-20090406 000874632 980__ $$aI:(DE-Juel1)IAS-6-20130828 000874632 980__ $$aI:(DE-Juel1)INM-10-20170113 000874632 980__ $$aAPC 000874632 980__ $$aUNRESTRICTED 000874632 981__ $$aI:(DE-Juel1)IAS-6-20130828