Department Colloquium: Tuesday April 11, 2023 - Novel algorithms and Statistical techniques in search of exoplanets

Event Date: 

Tuesday, April 11, 2023 - 3:45pm

Event Location: 

  • Broida 1640 and Zoom
  • Physics Department Colloquium

This event will be in-person and on Zoom.

Novel algorithms and Statistical techniques in search of exoplanets

Barak Zackay, Weizmann Institute

Exoplanet research is at the forefront of our understanding of solar system formation, evolution and of our place in the universe.

Data analysis is a main component of every astronomical survey. Novel data analysis tools, just like other types of technology, can substantially improve survey performance and reveal phenomena that we could not observe before.
I will show novel data analysis techniques that open new possibilities in exoplanet detection. I will demonstrate this on two of the main exoplanet detection channels:

1) Precision radial velocity - measuring the reflex motion of the star due to the planet's orbital motion, is the main channel that can provide mass measurements for transiting planets, as well as a main channel for discovering interesting planets for future followup missions. The precision of radial velocity measurements had plateaued in the last decade on approx 0.5-1 m/s precision. The main source of noise is the stellar variability itself. I will show a formalism that allows to completely bypass the barrier of stellar variability, and precisely measure the radial velocity, to the instrumental limit.

2) Transiting surveys - measuring the slight decrease in flux of the primary star due to the transit of the planet in front of it's host. The Kepler data, taken by the Kepler spacecraft is still our best data set for measuring the abundance of small planets with long orbital periods. I will show how methods developed for gravitational wave data analysis come into play for making a sensitive and reliable pipeline for detecting planets in the presence of red-noise, and preliminary results from running it on the Kepler data set.



Barak Zackay