Search

We present a novel natural language processing (NLP) approach to deriving plain English descriptors for science cases otherwise restricted by obfuscating technical terminology. We address the limitations of common radio galaxy morphology classifications by applying this approach. We experimentally derive a set of semantic tags for the Radio Galaxy Zoo EMU (Evolutionary Map of the Universe) project and the wider astronomical community. We collect 8486 plain English annotations of radio galaxy morphology, from which we derive a taxonomy of tags. The tags are plain English. The result is an extensible framework, which is more flexible, more easily communicated, and more sensitive to rare feature combinations, which are indescribable using the current framework of radio astronomy classifications. [ABSTRACT FROM AUTHOR]

The Evolutionary Map of the Universe (EMU) large-area radio continuum survey will detect tens of millions of radio galaxies, giving an opportunity for the detection of previously unknown classes of objects. To maximize the scientific value and make new discoveries, the analysis of these data will need to go beyond simple visual inspection. We propose the coarse-grained complexity, a simple scalar quantity relating to the minimum description length of an image that can be used to identify unusual structures. The complexity can be computed without reference to the broader sample or existing catalogue data, making the computation efficient on new surveys at very large scales (such as the full EMU survey). We apply our coarse-grained complexity measure to data from the EMU Pilot Survey to detect and confirm anomalous objects in this data set and produce an anomaly catalogue. Rather than work with existing catalogue data using a specific source detection algorithm, we perform a blind scan of the area, computing the complexity using a sliding square aperture. The effectiveness of the complexity measure for identifying anomalous objects is evaluated using crowd-sourced labels generated via the Zooniverse.org platform. We find that the complexity scan identifies unusual sources, such as odd radio circles, by partitioning on complexity. We achieve partitions where 5 per cent of the data is estimated to be 86 per cent complete, and 0.5 per cent is estimated to be 94 per cent pure, with respect to anomalies and use this to produce an anomaly catalogue. [ABSTRACT FROM AUTHOR]