?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Denoising+source+separation&rft.creator=S%C3%A4rel%C3%A4%2C+Mr+Jaakko&rft.creator=Valpola%2C+Dr+Harri&rft.subject=Statistical+Models&rft.subject=Machine+Learning&rft.subject=Neural+Nets&rft.subject=Artificial+Intelligence&rft.description=A+new+algorithmic+framework+called+denoising+source+separation+(DSS)+is+introduced.+The+main+benefit+of+this+framework+is+that+it+allows+for+easy+development+of+new+source+separation+algorithms+which+are+optimised+for+specific+problems.+In+this+framework%2C+source+separation+algorithms+are+constucted+around+denoising+procedures.+The+resulting+algorithms+can+range+from+almost+blind+to+highly+specialised+source+separation+algorithms.+Both+simple+linear+and+more+complex+nonlinear+or+adaptive+denoising+schemes+are%0Aconsidered.+Some+existing+independent+component+analysis+algorithms+are+reinterpreted+within+DSS+framework+and+new%2C+robust+blind+source+separation+algorithms+are+suggested.++Although+DSS+algorithms+need+not+be+explicitly+based+on+objective+functions%2C+there+is+often+an+implicit+objective+function+that+is+optimised.+The+exact+relation+between+the+denoising+procedure+and+the+objective+function+is+derived+and+a+useful+approximation+of+the+objective+function+is+presented.++In+the+experimental+section%2C+various+DSS+schemes++are+applied+extensively+to+artificial+data%2C+to+real+magnetoencephalograms+and+to+simulated+CDMA+mobile+network+signals.+Finally%2C+various+extensions+to+the+proposed+DSS+algorithms+are+considered.+These+include+nonlinear+observation+mappings%2C+hierarchical+models+and+overcomplete%2C++nonorthogonal+feature+spaces.+With+these+extensions%2C+DSS+appears+to+have+relevance+to+many+existing+models+of+neural+information+processing.%0A&rft.date=2004-02&rft.type=Preprint&rft.type=NonPeerReviewed&rft.format=application%2Fpdf&rft.identifier=http%3A%2F%2Fcogprints.org%2F3441%2F1%2FJS_DSS.pdf&rft.identifier=++S%C3%A4rel%C3%A4%2C+Mr+Jaakko+and+Valpola%2C+Dr+Harri++(2004)+Denoising+source+separation.++%5BPreprint%5D+++++&rft.relation=http%3A%2F%2Fcogprints.org%2F3441%2F