Artificial Intelligence finds six new pulsars in Arecibo data.

Image may be NSFW.
Clik here to view.© UBC

In the modern era of big data, many fields of astronomy are generating huge volumes, the analysis of which can sometimes be the limiting factor in research. Fortunately, powerful data-mining techniques have been developed by computer scientists, ready to be applied to various fields. In a recent paper, Wei Wei Zhu and Aaron Berndsen from the University of British Columbia plus collaborators present a novel artificial intelligence (AI) program that identifies pulsars from recent surveys using image pattern recognition with deep neural nets the PICS (Pulsar Image- based Classification System) AI.

Traditionally, new pulsars are identified from diagnostic images by humans (such as the new pulsars found in the LOFAR pilot surveys). The AI mimics these human experts and distinguishes pulsars from noise and interference by looking for patterns from candidate plots. Different from other pulsar selection programs which searched for expected patterns, the PICS AI is taught the salient features of different pulsars from a set of human-labeled candidates through machine learning.

The training candidates are collected from the Pulsar Arecibo L-band Feed Array Survey, which is carried out by a team including ASTRON's Jason Hessels and Joeri van Leeuwen. The information from each pulsar candidate is synthesized in four diagnostic plots, which consist of image data with up to thousands of pixels. The AI takes these data from each candidate as its input and uses thousands of such candidates to train its 9000 neurons. The deep neural networks in this AI system grant it superior ability in recognizing various types of pulsars as well as their harmonic signals.

The performance of this system can be improved over time as more training data are accumulated. This AI system has been integrated into the PALFA survey pipeline and has discovered six new pulsars to date!