Deep-learning based measurement of planetary radial velocities in the presence of stellar variability

TL;DR

This paper introduces a deep learning approach using neural networks to accurately measure small planetary radial velocities amidst stellar variability, demonstrating high precision recovery of injected signals in spectra.

Contribution

The study develops and compares neural network architectures, notably a multi-line CNN, to improve detection of tiny planetary signals in stellar spectra, advancing RV measurement techniques.

Findings

Multi-line CNN recovers 0.2 m/s semi-amplitude planets with 8.8% amplitude error

Achieves 0.7% error in period estimation for 50-day period planets

Demonstrates potential for detecting small planetary RVs with unprecedented accuracy

Abstract

We present a deep-learning based approach for measuring small planetary radial velocities in the presence of stellar variability. We use neural networks to reduce stellar RV jitter in three years of HARPS-N sun-as-a-star spectra. We develop and compare dimensionality-reduction and data splitting methods, as well as various neural network architectures including single line CNNs, an ensemble of single line CNNs, and a multi-line CNN. We inject planet-like RVs into the spectra and use the network to recover them. We find that the multi-line CNN is able to recover planets with 0.2 m/s semi-amplitude, 50 day period, with 8.8% error in the amplitude and 0.7% in the period. This approach shows promise for mitigating stellar RV variability and enabling the detection of small planetary RVs with unprecedented precision.