Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Spatio-Spectroscopic Representation Learning using Unsupervised Convolutional Long-Short Term Memory Networks

Kameswara Bharadwaj Mantha, Lucy Fortson, Ramanakumar Sankar, Claudia Scarlata, Chris Lintott, Sandor Kruk, Mike Walmsley, Hugh Dickinson, Karen Masters, Brooke Simmons, Rebecca Smethurst

TL;DR

A new unsupervised deep learning framework is demonstrated using Convolutional Long-Short Term Memory Network Autoencoders to encode generalized feature representations across both spatial and spectroscopic dimensions across both spatial and spectroscopic dimensions.

Abstract

Integral Field Spectroscopy (IFS) surveys offer a unique new landscape in which to learn in both spatial and spectroscopic dimensions and could help uncover previously unknown insights into galaxy evolution. In this work, we demonstrate a new unsupervised deep learning framework using Convolutional Long-Short Term Memory Network Autoencoders to encode generalized feature representations across both spatial and spectroscopic dimensions spanning $19$ optical emission lines (3800A $< λ<$ 8000A) among a sample of $\sim 9000$ galaxies from the MaNGA IFS survey. As a demonstrative exercise, we assess our model on a sample of $290$ Active Galactic Nuclei (AGN) and highlight scientifically interesting characteristics of some highly anomalous AGN.

Spatio-Spectroscopic Representation Learning using Unsupervised Convolutional Long-Short Term Memory Networks

TL;DR

A new unsupervised deep learning framework is demonstrated using Convolutional Long-Short Term Memory Network Autoencoders to encode generalized feature representations across both spatial and spectroscopic dimensions across both spatial and spectroscopic dimensions.

Abstract

Integral Field Spectroscopy (IFS) surveys offer a unique new landscape in which to learn in both spatial and spectroscopic dimensions and could help uncover previously unknown insights into galaxy evolution. In this work, we demonstrate a new unsupervised deep learning framework using Convolutional Long-Short Term Memory Network Autoencoders to encode generalized feature representations across both spatial and spectroscopic dimensions spanning optical emission lines (3800A 8000A) among a sample of galaxies from the MaNGA IFS survey. As a demonstrative exercise, we assess our model on a sample of Active Galactic Nuclei (AGN) and highlight scientifically interesting characteristics of some highly anomalous AGN.
Paper Structure (12 sections, 1 equation, 3 figures)

This paper contains 12 sections, 1 equation, 3 figures.

Table of Contents

  1. Introduction
  2. Data
  3. MaNGA Target Sample & IFS Data Cubes
  4. MaNGA AGN-Hosting Galaxies
  5. Deep Learning Model Framework
  6. Methodology
  7. Data Preparation
  8. Training Strategy
  9. Results & Discussion
  10. Latent Space & Anomaly Score Distribution
  11. Assessment of Highly Anomalous AGN & Nearest Neighbor Search
  12. Conclusions

Figures (3)

  • Figure 1: A conceptual overview of our 2DConvLSTM-vAE framework. For the 2DConvLSTM-AE architecture, the mean and variance layers are omitted.
  • Figure 2: Visualization of latent space embedding for our $\sim 9000$ galaxies extracted from the 2DConvLSTM-AE (left) and 2DConvLSTM-vAE (right) using a UMAP along with the inset plot showing the histogram of anomaly scores.
  • Figure 3: Example highly anomalous AGN galaxies (top panel) and example showcase of latent space nearest neighbour search for two anomalous AGN as query galaxies (bottom panel).