CASI: A Convolutional Neural Network Approach for Shell Identification

Abstract

We utilize techniques from deep learning to identify signatures of stellar feedback in simulated molecular clouds. Specifically, we implement a deep neural network with an architecture similar to U-Net and apply it to the problem of identifying wind-driven shells and bubbles using data from magneto-hydrodynamic simulations of turbulent molecular clouds with embedded stellar sources. The network is applied to two tasks, dense regression and segmentation, on two varieties of data, simulated density and synthetic 12 CO observations. Our Convolutional Approach for Shell Identification (CASI) is able to obtain a true positive rate greater than 90\%, while maintaining a false positive rate of 1\%, on two segmentation tasks and also performs well on related regression tasks. The source code for CASI is available on GitLab.

References in corpus (45)

• Adam: A Method for Stochastic Optimization
• An overview of gradient descent optimization algorithms
• Road Extraction by Deep Residual U-Net
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Galaxy Zoo : Morphologies derived from visual inspection of galaxies from the Sloan Digital Sky Survey
• A guide to convolution arithmetic for deep learning
• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• Don't Decay the Learning Rate, Increase the Batch Size
• On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima
• AnatomyNet: Deep Learning for Fast and Fully Automated Whole-volume Segmentation of Head and Neck Anatomy
• The mass function of dense molecular cores and the origin of the IMF
• Radiation-Hydrodynamic Simulations of Collapse and Fragmentation in Massive Protostellar Cores
• Outflow Feedback Regulated Massive Star Formation in Parsec-Scale Cluster Forming Clumps
• Revisiting Small Batch Training for Deep Neural Networks
• Recent Advances in Convolutional Neural Networks
• The COMPLETE Survey of Star-Forming Regions: Phase I Data
• Inefficient star formation through turbulence, magnetic fields and feedback
• CMA-ES for Hyperparameter Optimization of Deep Neural Networks
• The Milky Way Project First Data Release: A Bubblier Galactic Disk
• CMU DeepLens: Deep Learning For Automatic Image-based Galaxy-Galaxy Strong Lens Finding
• The COMPLETE Survey of Outflows in Perseus
• A Bubbling Nearby Molecular Cloud: COMPLETE Shells in Perseus
• A multi-scale filament extraction method: getfilaments
• The Origin and Universality of the Stellar Initial Mass Function
• The Turbulent Origin of Outflow and Spin Misalignment in Multiple Star Systems
• Impact of Protostellar Outflows on Turbulence and Star Formation Efficiency in Magnetized Dense Cores
• Deep Learning Identifies High-z Galaxies in a Central Blue Nugget Phase in a Characteristic Mass Range
• Bubble size statistics during reionization from 21-cm tomography
• Radiation-Hydrodynamic Simulations of Protostellar Outflows: Synthetic Observations and Data Comparisons
• Residual Connections Encourage Iterative Inference
• Time Varying Dynamical Star Formation Rate
• The Milky Way Project: Leveraging Citizen Science and Machine Learning to Detect Interstellar Bubbles
• Outflows and Bubbles in Taurus: Star-formation Feedback Sufficient to Maintain Turbulence
• Impact of Winds from Intermediate-Mass Stars on Molecular Cloud Structure and Turbulence
• Molecular Outflows Identified in the FCRAO CO Survey of the Taurus Molecular Cloud
• Detecting solar system objects with convolutional neural networks
• Outflow Driven Cavities: Numerical Simulations of Intermediaries of Protostellar Turbulence
• Catalogue of far-infrared loops in the Galaxy
• Classifying structures in the ISM with Support Vector Machines: the G16.05-0.57 supernova remnant
• RUN:Residual U-Net for Computer-Aided Detection of Pulmonary Nodules without Candidate Selection
• Assessing the Performance of a Machine Learning Algorithm in Identifying Bubbles in Dust Emission
• Classifying the formation processes of S0 galaxies using Convolutional Neural Networks
• Automatic Detection of Expanding HI Shells Using Artificial Neural Networks
• Assessing the Impact of Astrochemistry on Molecular Cloud Turbulence Statistics
• 3D Convolutional Neural Networks for Classification of Functional Connectomes

Cited by in corpus (16)

• PHANGS-JWST First Results: A statistical view on bubble evolution in NGC628
• Zooming in on Individual Star Formation: Low- and High-mass Stars
• FEEDBACK from the NGC7538 HII region
• A Morphological Classification of 18190 Molecular Clouds Identified in ${}^{12}$CO Data from the MWISP Survey
• Application of Convolutional Neural Networks to Identify Stellar Feedback Bubbles in CO Emission
• Denoising Diffusion Probabilistic Models to Predict the Density of Molecular Clouds
• Application of Convolutional Neural Networks to Identify Protostellar Outflows in CO Emission
• Tracking down the origin of superbubbles and supergiant shells in the Magellanic Clouds with Minkowski tensor analysis
• PGNets: Planet mass prediction using convolutional neural networks for radio continuum observations of protoplanetary disks
• ML-LBM: Machine Learning Aided Flow Simulation in Porous Media
• A Census of Outflow to Magnetic Field Orientations in Nearby Molecular Clouds
• Application of Convolutional Neural Networks to Predict Magnetic Fields Directions in Turbulent Clouds
• A Census of Protostellar Outflows in Nearby Molecular Clouds
• MAGIC: Microlensing Analysis Guided by Intelligent Computation
• CLOVER: Convnet Line-fitting Of Velocities in Emission-line Regions
• A new way to constrain the densities of intra-group medium in groups of galaxies with convolutional neural networks