BEGIN:VCALENDAR VERSION:2.0 X-WR-CALNAME:EventsCalendar PRODID:-//hacksw/handcal//NONSGML v1.0//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York LAST-MODIFIED:20240422T053451Z TZURL:https://www.tzurl.org/zoneinfo-outlook/America/New_York X-LIC-LOCATION:America/New_York BEGIN:DAYLIGHT TZNAME:EDT TZOFFSETFROM:-0500 TZOFFSETTO:-0400 DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT BEGIN:STANDARD TZNAME:EST TZOFFSETFROM:-0400 TZOFFSETTO:-0500 DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT CATEGORIES:College of Engineering,Lectures and Seminars,Thesis/Dissertation s DESCRIPTION:Topic: Modeling Spin-Dependent Detectability in Gravitational-W ave Astronomy with a Calibrated Normalizing Flow Abstract: Population inference from gravitational-wave catalogs requires an accurate selection function, the probability that a source with given parameters i s detected, because errors in this correction propagate directly into the inferred astrophysical distributions. The standard semianalytic approach o f Finn and Chernoff estimates detectability from the leading-order post-Ne wtonian amplitude, which depends on chirp mass, luminosity distance, and o rientation but carries no dependence on component spin. Real waveforms are not spin-blind: aligned spin modifies the inspiral through spin-orbit cou pling, delays merger via the orbital hang-up, and raises the accumulated s ignal-to-noise ratio. A spin-blind selection function therefore misreprese nts the detectability of spinning binaries and the sensitive volume availa ble to spinning populations. This thesis quantifies that bias with a calib rated conditional normalizing flow trained on a large synthetic population of binary black hole signals, generated with a full precessing, higher-ha rmonic waveform model for the Advanced LIGO-Virgo network at design sensit ivity. Rather than classifying detection at a fixed threshold, the flow mo dels the full conditional signal-to-noise distribution and remains evaluab le at any threshold. Benchmarked against the Finn-Chernoff baseline, the f low recovers a strong dependence of sensitive volume on effective spin, sp anning a factor of roughly 2.6 between strongly anti-aligned and strongly aligned systems, whereas the baseline stays spin-independent by constructi on. This discrepancy is a spin-selection bias that must be accounted for i n spin-population inference as catalogs continue to grow. Advisor(s): Dr. Sarah Caudill, Department of Physics, (scaudill@umassd.edu) Committee memb ers: Dr. Robert Fisher, Department of Physics and Dr. Scott Field, Dep artment of Mathematics Note: All PHY Graduate Students are encouraged to a ttend. \nEvent page: https://www.umassd.edu/events/cms/20260727-physics- master-of-science-thesis-defense-by-sara-gholamhoseinian-.php\nEvent link: https://umassd.zoom.us/j/97464617175?pwd=1sGVbiZIj8rZZtEWLohylHQtoZQlt1.1 X-ALT-DESC;FMTTYPE=text/html:

Topic: Modeling Spin-Dependent Detectability in Gravitational-Wave Astronomy with a Calibrated Normalizin g Flow

Abstract:

Population inference from gravitational-wave catalogs requires an accurate selection function\, the probability that a source with given parameters is detected\, because erro rs in this correction propagate directly into the inferred astrophysical d istributions. The standard semianalytic approach of Finn and Chernoff esti mates detectability from the leading-order post-Newtonian amplitude\, whic h depends on chirp mass\, luminosity distance\, and orientation but carrie s no dependence on component spin. Real waveforms are not spin-blind: alig ned spin modifies the inspiral through spin-orbit coupling\, delays merger via the orbital hang-up\, and raises the accumulated signal-to-noise rati o. A spin-blind selection function therefore misrepresents the detectabili ty of spinning binaries and the sensitive volume available to spinning pop ulations. This thesis quantifies that bias with a calibrated conditional n ormalizing flow trained on a large synthetic population of binary black ho le signals\, generated with a full precessing\, higher-harmonic waveform m odel for the Advanced LIGO-Virgo network at design sensitivity. Rather tha n classifying detection at a fixed threshold\, the flow models the full co nditional signal-to-noise distribution and remains evaluable at any thresh old. Benchmarked against the Finn-Chernoff baseline\, the flow recovers a strong dependence of sensitive volume on effective spin\, spanning a facto r of roughly 2.6 between strongly anti-aligned and strongly aligned system s\, whereas the baseline stays spin-independent by construction. This disc repancy is a spin-selection bias that must be accounted for in spin-popula tion inference as catalogs continue to grow.

Advisor(s): Dr. Sarah Caudill\, Department of Physics\, (scaudill@umassd.edu)

Committee members: Dr. Robert Fisher\, Department of Physics and Dr . Scott Field\, Department of Mathematics

Note: All PHY Graduate Students are encouraged to attend.

Event page: https://www.umassd.edu/events/cms/20260727-physics-master-o f-science-thesis-defense-by-sara-gholamhoseinian-.php
Event link: < a href="https://umassd.zoom.us/j/97464617175?pwd=1sGVbiZIj8rZZtEWLohylHQto ZQlt1.1">https://umassd.zoom.us/j/97464617175?pwd=1sGVbiZIj8rZZtEWLohylHQt oZQlt1.1

DTSTAMP:20260708T234959 DTSTART;TZID=America/New_York:20260727T120000 DTEND;TZID=America/New_York:20260727T140000 LOCATION:LIB 314 SUMMARY;LANGUAGE=en-us:Physics Master of Science Thesis Defense by Sara Gho lamhoseinian UID:477d214e6b3263b6189653a86185a1fd@www.umassd.edu END:VEVENT END:VCALENDAR