BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Armenak Petrosyan (Georgia institute of Technology) DTSTART:20220430T140000Z DTEND:20220430T150000Z DTSTAMP:20260423T052837Z UID:YMC/35 DESCRIPTION:Title: Har monic Analysis and Approximation Theory techniques in Machine Learning \nby Armenak Petrosyan (Georgia institute of Technology) as part of Yereva n Mathematical Colloquium\n\n\nAbstract\nIn this talk\, we will discuss th e existing work and possible future promising areas of interest in the int erplay between Machine Learning\, Harmonic Analysis\, and Approximation Th eory. The presentation will cover an overview of the research progress in the crossroads of these fields\, as well as my own contributions in the ar ea of neural network approximations.\n\nArtificial neural networks have ga ined widespread adoption as a powerful tool for various machine learning t asks in recent years. \nOne challenge we will address in this talk is to f ind small size shallow neural networks that can be trained algorithmically and which achieve guaranteed approximation speed and precision. To mainta in the small size we apply penalties on the weights of the network. We sho w that under minimal requirements\, all local minima of the resulting prob lem are well behaved and possess a desirable small size without sacrificin g precision.\n\nTalk host: Michael Poghosyan (YSU)\n LOCATION:https://researchseminars.org/talk/YMC/35/ END:VEVENT END:VCALENDAR