BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Pablo Suárez Serrato (UNAM - Mexico) DTSTART:20240426T160000Z DTEND:20240426T170000Z DTSTAMP:20260423T022915Z UID:GEOTOP-A/67 DESCRIPTION:Title: Topics in Geometric Learning\nby Pablo Suárez Serrato (UNAM - Mexic o) as part of GEOTOP-A seminar\n\n\nAbstract\nSimilarly to the growth of A pplied Topology\, the uses and applications of Geometry are now expanding into scientific\, computational\, and engineering domains. First\, we'll r eview the recent history of this expanding Applied Geometry area. I'll men tion several collaborations. Developing and implementing algorithms inspir ed by the marked length spectrum that classifies complex networks (with El iassi-Rad and Torres) and analyzing digital images using a variant of curv e-shortening flow (with Velazquez Richards). As well as a definition I pro posed of a global convolution on manifolds of arbitrary topology\, relevan t for deep learning on manifolds. Furthermore\, I'll present our joint wor k with Evangelista and Ruiz Pantaleón on computational Poisson geometry. This work includes a practical application in learning symbolic expression s of Hamiltonian systems. We've developed and released two Python packages that are instrumental in this process. These packages enable symbolic and numerical computations of objects in Poisson geometry\, and they're compa tible with the deep learning frameworks NumPy\, TensorFlow\, and PyTorch. We've utilized these packages to train neural networks\, particularly hybr ids with CNN and LSTM components\, that learn symbolic expressions of Hami ltonian vector fields. I'll present a tutorial on our computational Poisso n Geometry modules if time allows.\n LOCATION:https://researchseminars.org/talk/GEOTOP-A/67/ END:VEVENT END:VCALENDAR