Log[2]: Assembling Background, Purpose, and Implication
October 27, 2021
Background
Kepler Space Telescope
The Kepler mission was designed to survey a region of the Milky Way galaxy to detect and characterize Earth-size and smaller planets in or near the habitable zone by using the transit method of planetary detection. This was accomplished by observing changes in the brightness of stars in the same patch of sky for 4 years between May 2009 and May 2013.
Shazam
Shazam analyses the captured sound of a song and finds the name of the song. based on an acoustic fingerprint in a database of millions of songs
Purpose & Implications
Influx of Data
Hubble Space Telescope produced approximately 3 GB per day. James Webb Space Telescope (JWST) expected to produce approximately 57.5 GB per day (Beichman et al. 2014). Square Kilometer Array, which will be online in 2027, is predicted to produce on the order of 109 GB per day It will generate data streams far beyond the total Internet traffic worldwide. Impossible for humans to pick out and sort through.
Hidden Features
As shown by Shazam, many features might not be recognizable by human researchers
Incomplete & Wasted Data
As shown by Shazam, it is hard for humans or models to recognize incomplete datasets. However, specifically-trained neural networks can do that.
Machine Learning Application in Kepler Light Curves
Not much ML research has been applied on Kepler datasets. Many related works in supernovae ML projects.