Special Talk

• Special Talks

  Speaker: Digvijay (Jay) Wadekar (JHU)
  Title: Best of Both Worlds: Integrating Principled Matched-Filtering GW Searches with AI/ML Tools​​​​​​​
  Abstract: In the infinite data and compute limit, machine learning (ML) methods can be optimal, however this idealistic situation is not often realized in practice. On the other hand, principled data-analysis methods are robust, but they make simplistic assumptions (e.g., the noise is roughly Gaussian). I will present how ML algorithms can enhance matched-filtering gravitational wave pipelines by: (i) generating optimal template banks (ii) weighting templates to downplay unphysical binary configurations (iii) mitigating non-Gaussian noise. Incorporating these advancements in the IAS search pipeline, I will present new detections of black holes in the astrophysically significant pair-instability mass gap and IMBH mass ranges.

  Speaker: Russell Ryan (STScI)​​​​​​​
  Title: The Roman Supernova Cosmology Project Infrastructure Team
  Abstract: One of the central science requirements for the Roman Space Telescope is to measure and characterize the expansion history of the Universe by discovering and monitoring SNe Ia for z<2.5. To this end, it will conduct the High-Latitude Time-Domain Survey, which will cover ~18.3 deg2 in the RZYJHF bands with ~5 deg2 of slitless spectrosocpy distributed over a ~2 year baseline. To support this survey and ensure the science return, NASA commissioned the Supernova Project Infrastructure Team, which involves numerous STScI members. I will discuss the work of Infrastructure Team, particularly focusing on the role of STScI researchers. Although our work spans a wide range from software development and calibration, the largest component involves the slitless spectroscopic survey. I am the spectroscopy lead for the Infrastructure Team, which is developing a pipeline for pixels-to-cosmology with functionality to mitigate contamination, determine stellar-population parameters of the host galaxies, extract and combine one-dimensional spectra, and measure high-level properties. Our tools are open source and deployable locally or in the cloud with thoughts to the Roman Nexus. Based on the current survey definitions, the program is expected to begin mid 2028 and spectroscopically observe ~10 transients/day (~1 SNIa/day). This spectroscopic training sample is foundational to bootstrap the photometric-only sample by providing spectral typing, light-curve standarization, redshift, and SED templates.