Spatial distribution of the large-scale structure: An unsupervised search for parity violation

We use machine learning methods to search for parity violations in the large-scale structure (LSS) of the Universe, motivated by recent claims of chirality detection using the 4-point correlation function (4PCF), which would suggest new physics during the epoch of inflation. This work seeks to reproduce these claims using methods originating from high energy collider analyses. Our machine learning methods optimize some underlying parity odd function of the data, and use it to evaluate the parity odd fraction. We demonstrate the effectiveness and suitability of these methods and then apply them to the Baryon Oscillation Spectroscopic Survey (BOSS) catalog. No strong evidence for parity violation is detected.