BEGIN:VCALENDAR METHOD:REQUEST PRODID:Microsoft Exchange Server 2010 VERSION:2.0 BEGIN:VTIMEZONE TZID:Eastern Standard Time BEGIN:STANDARD DTSTART:16010101T020000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;INTERVAL=1;BYDAY=1SU;BYMONTH=11 END:STANDARD BEGIN:DAYLIGHT DTSTART:16010101T020000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;INTERVAL=1;BYDAY=2SU;BYMONTH=3 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT ORGANIZER;CN=Michael McDonald:mailto:michael.mcdonald@fsu.edu DESCRIPTION;LANGUAGE=en-US:"PGINN -- Population Genetics Informed Neural Ne tworks"\n\nPeter Beerli\nDepartment of Scientific Computing\nFlorida State University (FSU)\n\nPlease feel free to forward/share this invitation wit h other groups/disciplines that might be interested in this talk/topic. Al l are welcome to attend.\n\nNOTE: In-person attendance is requested in our 499 Dirac Science Library (DSL) Seminar Room. Zoom access is intended for external (non-departmental) participants only.\n\nhttps://fsu.zoom.us/j/9 4273595552\n\nMeeting # 942 7359 5552\n\n\nWednesday\, Nov 4\, 2026\, Sche dule:\n\n\n *\n3:00 to 3:30 PM Eastern Time (US and Canada)\n *\nā˜• Nespresso & Teatime - 417 DSL Commons\n\n *\n3:30 to 4:30 PM Eastern Time (US and Canada)\n *\nšŸ•Ÿ Colloquium - 499 DSL Seminar Room\n\n\n\nAb stract:\n\nNeural networks trained on simulated genealogies estimate popul ation-genetic parameters quickly\, but their loss ignores the model that g enerated the data. Following physics-informed neural networks\, we introdu ce PGINN (Population-Genetics-Informed Neural Networks)\, which add to the loss a term built from the coalescent rate along each genealogy. We study one population under a completed hard sweep\, with 10 independent data se ts per comparison. With log-scaled input\, the local rate divergence impro ves the estimate of the population-scaled selection coefficient s in every regime: it removes 7 to 12% of the squared error in log(1+s) of a network trained without it. Penalties exact for the coalescent process give littl e gain. The local rate divergence helps through its asymmetry: it penalize s an underestimate of strong selection and counters the shrinkage of netwo rks trained with mean squared error (MSE). With known genealogies\, maximu m likelihood is more accurate still. The gain carries over to genealogies estimated by the software MIGRATE from simulated recombining chromosomes: networks with the local rate divergence detect sweeps with s=50 to 200 mor e often than MSE in 9 of 10 training runs\, and both detect weak sweeps mo re than twice as often as a scan for low ƎĖœ. At the Adh locus of Drosoph ila melanogaster\, estimates of s are consistent with neutrality. The pena lty needs only coalescent rates\, so changing population size\, structure or migration can be handled the same way. Where likelihood methods are una vailable or too slow\, PGINN offers a simple way to bring population-genet ic theory into network training.\n\n\nAdditional colloquium details can be found here:\nhttps://www.sc.fsu.edu/news-and-events/colloquium/colloquium -with-peter-beerli-2026-11-04\n\n\nšŸŽ¦ Colloquium recordings will be made available here:\ nhttps://www.sc.fsu.edu/news-and-events/seminars\n UID:040000008200E00074C5B7101A82E008000000008683A842006BDD01000000000000000 0100000001B3921BC781E7846A94EB5CEB582F718 SUMMARY;LANGUAGE=en-US:Scientific Computing Colloquium with Peter Beerli DTSTART;TZID=Eastern Standard Time:20261104T153000 DTEND;TZID=Eastern Standard Time:20261104T162500 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20261102T192755Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:1 LOCATION;LANGUAGE=en-US:https://fsu.zoom.us/j/94273595552 X-MICROSOFT-CDO-APPT-SEQUENCE:1 X-MICROSOFT-CDO-OWNERAPPTID:2125141690 X-MICROSOFT-CDO-BUSYSTATUS:TENTATIVE X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE X-MICROSOFT-LOCATIONDISPLAYNAME:https://fsu.zoom.us/j/94273595552 X-MICROSOFT-LOCATIONSOURCE:None X-MICROSOFT-LOCATIONS:[{"DisplayName":"https://fsu.zoom.us/j/94273595552"\, "LocationAnnotation":""\,"LocationUri":""\,"LocationStreet":""\,"LocationC ity":""\,"LocationState":""\,"LocationCountry":""\,"LocationPostalCode":"" \,"LocationFullAddress":""}] BEGIN:VALARM DESCRIPTION:REMINDER TRIGGER;RELATED=START:-PT5M ACTION:DISPLAY END:VALARM END:VEVENT END:VCALENDAR