PGNets: Planet mass prediction using convolutional neural networks for radio continuum observations of protoplanetary disks

Abstract

We developed Convolutional Neural Networks (CNNs) to rapidly and directly infer the planet mass from radio dust continuum images. Substructures induced by young planets in protoplanetary disks can be used to infer the potential young planets' properties. Hydrodynamical simulations have been used to study the relationships between the planet's properties and these disk features. However, these attempts either fine-tuned numerical simulations to fit one protoplanetary disk at a time, which was time-consuming, or azimuthally averaged simulation results to derive some linear relationships between the gap width/depth and the planet mass, which lost information on asymmetric features in disks. To cope with these disadvantages, we developed Planet Gap neural Networks (PGNets) to infer the planet mass from 2D images. We first fit the gridded data in Zhang et al. (2018) as a classification problem. Then, we quadrupled the data set by running additional simulations with near-randomly sampled parameters, and derived the planet mass and disk viscosity together as a regression problem. The classification approach can reach an accuracy of 92\%, whereas the regression approach can reach 1$\sigma$ as 0.16 dex for planet mass and 0.23 dex for disk viscosity. We can reproduce the degeneracy scaling $\alpha$ $\propto$ $M_p^3$ found in the linear fitting method, which means that the CNN method can even be used to find degeneracy relationship. The gradient-weighted class activation mapping effectively confirms that PGNets use proper disk features to constrain the planet mass. We provide programs for PGNets and the traditional fitting method from Zhang et al. (2018), and discuss each method's advantages and disadvantages.

12 pages, 7 figures, accepted to MNRAS

References in corpus (29)

• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• Two accreting protoplanets around the young star PDS 70
• Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks
• On the corotation torque in a radiatively inefficient disk
• Dust flow in gas disks in the presence of embedded planets
• Mass Estimates of a Giant Planet in a Protoplanetary Disk from the Gap Structures
• One Solution to the Mass Budget Problem for Planet Formation: Optically Thick Disks with Dust Scattering
• Submillimeter emission associated with candidate protoplanets
• Type I planetary migration in a self-gravitating disk
• The newborn planet population emerging from ring-like structures in discs
• The anomalously low (sub)millimeter spectral indices of some protoplanetary disks may be explained by dust self-scattering
• The Ophiuchus DIsc Survey Employing ALMA (ODISEA)-III: the evolution of substructures in massive discs at 3-5 au resolution
• An opening criterion for dust gaps in protoplanetary discs
• Keck/NIRC2 $L$'-Band Imaging of Jovian-Mass Accreting Protoplanets around PDS 70
• The Disc Miner I: A statistical framework to detect and quantify kinematical perturbations driven by young planets in discs
• The widest H$\alpha$ survey of accreting protoplanets around nearby transition disks
• Separating extended disc features from the protoplanet in PDS 70 using VLT/SINFONI
• The effects of disk self-gravity and radiative cooling on the formation of gaps and spirals by young planets
• A super-resolution analysis of the DSHARP survey: Substructure is common in the inner 30 au
• Revealing signatures of planets migrating in protoplanetary discs with ALMA multi-wavelength observations
• Constraining Protoplanetary Disk Accretion and Young Planets Using ALMA Kinematic Observations
• Using Deep Neural Networks to compute the mass of forming planets
• Application of Convolutional Neural Networks to Identify Stellar Feedback Bubbles in CO Emission
• Morphological signatures induced by dust back reaction in discs with an embedded planet
• Migrating Low-Mass Planets in Inviscid Dusty Protoplanetary Discs
• Application of Convolutional Neural Networks to Identify Protostellar Outflows in CO Emission
• Hiding Signatures of Gravitational Instability in Protoplanetary Discs with Planets
• A Machine Learning model to infer planet masses from gaps observed in protoplanetary disks
• DPNNet-2.0 Part I: Finding hidden planets from simulated images of protoplanetary disk gaps

Cited by in corpus (8)

• PPDONet: Deep Operator Networks for Fast Prediction of Steady-State Solutions in Disk-Planet Systems
• Substructures in Compact Disks of the Taurus Star-forming Region
• Denoising Diffusion Probabilistic Models to Predict the Density of Molecular Clouds
• DBNets: A publicly available deep learning tool to measure the masses of young planets in dusty protoplanetary discs
• Using Bayesian Deep Learning to infer Planet Mass from Gaps in Protoplanetary Disks
• DBNets2.0: simulation-based inference for planet-induced dust substructures in protoplanetary discs
• Kinematic Evidence of an Embedded Protoplanet in HD 142666 Identified by Machine Learning
• exoALMA XXIII. Estimating Disk and Planet Properties from Dust Morphologies with DBNets2.0