Perlustration of error surfaces for nonlinear stochastic gradient descent algorithms

We attempt to explain in more detail the performance of several novel algorithms for nonlinear neural adaptive filtering. Weight trajectories together with the error surface give a clear understandable representation of the family of least mean square (LMS) based, nonlinear gradient descent (NGD), s...

Full description

Saved in:
Bibliographic Details
Main Authors: Hanna, A.I., Krcmar, I.R., Mandic, D.P.
Format: Conference Proceeding
Language:English
Subjects:
Algorithm design and analysis
Backpropagation algorithms
Filters
Information systems
Least squares approximation
Monte Carlo methods
Performance analysis
Signal processing algorithms
Stochastic processes
Visualization
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page 16
container_issue
container_start_page 11
container_title
container_volume
creator Hanna, A.I.
Krcmar, I.R.
Mandic, D.P.
description We attempt to explain in more detail the performance of several novel algorithms for nonlinear neural adaptive filtering. Weight trajectories together with the error surface give a clear understandable representation of the family of least mean square (LMS) based, nonlinear gradient descent (NGD), search-then-converge (STC) learning algorithms and the real-time recurrent learning (RTRL) algorithm. Performance is measured on prediction of coloured and nonlinear input. The results are an alternative qualitative representation of different qualitative performance measures for the analysed algorithms. Error surfaces and the adjacent instantaneous prediction errors support the analysis.
doi_str_mv 10.1109/NEUREL.2002.1057958
format conference_proceeding
fullrecord <record><control><sourceid>ieee_6IE</sourceid><recordid>TN_cdi_ieee_primary_1057958</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>1057958</ieee_id><sourcerecordid>1057958</sourcerecordid><originalsourceid>FETCH-LOGICAL-i175t-fda3d65008a646230e81da11380d8f6669279da45fa8f43fe9fa9500cafb2cd03</originalsourceid><addsrcrecordid>eNotj9tKAzEYhAMiqLVP0Ju8wK5_NpvTpZT1AEsVaa_Lbw5tZLsrSXrh27ti5-ZjhmFgCFkxqBkD87Dpdh9dXzcATc1AKCP0FbkDpYErYbi5Icucv2AWN0K14pZs330azrkkLHEa6RSoT2lKNJ9TQOszDbMZp3GIo8c5LpM9Yi7R0kNCF_1YqPPZ_hGHw5RiOZ7yPbkOOGS_vHBBdk_ddv1S9W_Pr-vHvopMiVIFh9xJAaBRtrLh4DVzyBjX4HSQUppGGYetCKhDy4M3Ac1ctxg-G-uAL8jqfzd67_ffKZ4w_ewvx_kvZqhRLg</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>Perlustration of error surfaces for nonlinear stochastic gradient descent algorithms</title><source>IEEE Electronic Library (IEL) Conference Proceedings</source><creator>Hanna, A.I. ; Krcmar, I.R. ; Mandic, D.P.</creator><creatorcontrib>Hanna, A.I. ; Krcmar, I.R. ; Mandic, D.P.</creatorcontrib><description>We attempt to explain in more detail the performance of several novel algorithms for nonlinear neural adaptive filtering. Weight trajectories together with the error surface give a clear understandable representation of the family of least mean square (LMS) based, nonlinear gradient descent (NGD), search-then-converge (STC) learning algorithms and the real-time recurrent learning (RTRL) algorithm. Performance is measured on prediction of coloured and nonlinear input. The results are an alternative qualitative representation of different qualitative performance measures for the analysed algorithms. Error surfaces and the adjacent instantaneous prediction errors support the analysis.</description><identifier>ISBN: 0780375939</identifier><identifier>ISBN: 9780780375932</identifier><identifier>DOI: 10.1109/NEUREL.2002.1057958</identifier><language>eng</language><publisher>IEEE</publisher><subject>Algorithm design and analysis ; Backpropagation algorithms ; Filters ; Information systems ; Least squares approximation ; Monte Carlo methods ; Performance analysis ; Signal processing algorithms ; Stochastic processes ; Visualization</subject><ispartof>6th Seminar on Neural Network Applications in Electrical Engineering, 2002, p.11-16</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/1057958$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,776,780,785,786,2052,4036,4037,27902,54895</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/1057958$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Hanna, A.I.</creatorcontrib><creatorcontrib>Krcmar, I.R.</creatorcontrib><creatorcontrib>Mandic, D.P.</creatorcontrib><title>Perlustration of error surfaces for nonlinear stochastic gradient descent algorithms</title><title>6th Seminar on Neural Network Applications in Electrical Engineering</title><addtitle>NEUREL</addtitle><description>We attempt to explain in more detail the performance of several novel algorithms for nonlinear neural adaptive filtering. Weight trajectories together with the error surface give a clear understandable representation of the family of least mean square (LMS) based, nonlinear gradient descent (NGD), search-then-converge (STC) learning algorithms and the real-time recurrent learning (RTRL) algorithm. Performance is measured on prediction of coloured and nonlinear input. The results are an alternative qualitative representation of different qualitative performance measures for the analysed algorithms. Error surfaces and the adjacent instantaneous prediction errors support the analysis.</description><subject>Algorithm design and analysis</subject><subject>Backpropagation algorithms</subject><subject>Filters</subject><subject>Information systems</subject><subject>Least squares approximation</subject><subject>Monte Carlo methods</subject><subject>Performance analysis</subject><subject>Signal processing algorithms</subject><subject>Stochastic processes</subject><subject>Visualization</subject><isbn>0780375939</isbn><isbn>9780780375932</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2002</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNotj9tKAzEYhAMiqLVP0Ju8wK5_NpvTpZT1AEsVaa_Lbw5tZLsrSXrh27ti5-ZjhmFgCFkxqBkD87Dpdh9dXzcATc1AKCP0FbkDpYErYbi5Icucv2AWN0K14pZs330azrkkLHEa6RSoT2lKNJ9TQOszDbMZp3GIo8c5LpM9Yi7R0kNCF_1YqPPZ_hGHw5RiOZ7yPbkOOGS_vHBBdk_ddv1S9W_Pr-vHvopMiVIFh9xJAaBRtrLh4DVzyBjX4HSQUppGGYetCKhDy4M3Ac1ctxg-G-uAL8jqfzd67_ffKZ4w_ewvx_kvZqhRLg</recordid><startdate>2002</startdate><enddate>2002</enddate><creator>Hanna, A.I.</creator><creator>Krcmar, I.R.</creator><creator>Mandic, D.P.</creator><general>IEEE</general><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>2002</creationdate><title>Perlustration of error surfaces for nonlinear stochastic gradient descent algorithms</title><author>Hanna, A.I. ; Krcmar, I.R. ; Mandic, D.P.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i175t-fda3d65008a646230e81da11380d8f6669279da45fa8f43fe9fa9500cafb2cd03</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2002</creationdate><topic>Algorithm design and analysis</topic><topic>Backpropagation algorithms</topic><topic>Filters</topic><topic>Information systems</topic><topic>Least squares approximation</topic><topic>Monte Carlo methods</topic><topic>Performance analysis</topic><topic>Signal processing algorithms</topic><topic>Stochastic processes</topic><topic>Visualization</topic><toplevel>online_resources</toplevel><creatorcontrib>Hanna, A.I.</creatorcontrib><creatorcontrib>Krcmar, I.R.</creatorcontrib><creatorcontrib>Mandic, D.P.</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Hanna, A.I.</au><au>Krcmar, I.R.</au><au>Mandic, D.P.</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Perlustration of error surfaces for nonlinear stochastic gradient descent algorithms</atitle><btitle>6th Seminar on Neural Network Applications in Electrical Engineering</btitle><stitle>NEUREL</stitle><date>2002</date><risdate>2002</risdate><spage>11</spage><epage>16</epage><pages>11-16</pages><isbn>0780375939</isbn><isbn>9780780375932</isbn><abstract>We attempt to explain in more detail the performance of several novel algorithms for nonlinear neural adaptive filtering. Weight trajectories together with the error surface give a clear understandable representation of the family of least mean square (LMS) based, nonlinear gradient descent (NGD), search-then-converge (STC) learning algorithms and the real-time recurrent learning (RTRL) algorithm. Performance is measured on prediction of coloured and nonlinear input. The results are an alternative qualitative representation of different qualitative performance measures for the analysed algorithms. Error surfaces and the adjacent instantaneous prediction errors support the analysis.</abstract><pub>IEEE</pub><doi>10.1109/NEUREL.2002.1057958</doi><tpages>6</tpages></addata></record>
fulltext fulltext_linktorsrc
identifier ISBN: 0780375939
ispartof 6th Seminar on Neural Network Applications in Electrical Engineering, 2002, p.11-16
issn
language eng
recordid cdi_ieee_primary_1057958
source IEEE Electronic Library (IEL) Conference Proceedings
subjects Algorithm design and analysis
Backpropagation algorithms
Filters
Information systems
Least squares approximation
Monte Carlo methods
Performance analysis
Signal processing algorithms
Stochastic processes
Visualization
title Perlustration of error surfaces for nonlinear stochastic gradient descent algorithms
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-31T03%3A00%3A52IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_6IE&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Perlustration%20of%20error%20surfaces%20for%20nonlinear%20stochastic%20gradient%20descent%20algorithms&rft.btitle=6th%20Seminar%20on%20Neural%20Network%20Applications%20in%20Electrical%20Engineering&rft.au=Hanna,%20A.I.&rft.date=2002&rft.spage=11&rft.epage=16&rft.pages=11-16&rft.isbn=0780375939&rft.isbn_list=9780780375932&rft_id=info:doi/10.1109/NEUREL.2002.1057958&rft_dat=%3Cieee_6IE%3E1057958%3C/ieee_6IE%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i175t-fda3d65008a646230e81da11380d8f6669279da45fa8f43fe9fa9500cafb2cd03%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=1057958&rfr_iscdi=true