Repeating versus Nonrepeating Fast Radio Bursts: A Deep Learning Approach to Morphological Characterization

We present a deep learning approach to classify fast radio bursts (FRBs) based purely on morphology as encoded on recorded dynamic spectrum from CHIME/FRB Catalog 2. We implemented transfer learning with a pretrained ConvNext architecture, exploiting its powerful feature extraction ability. ConvNext was adapted to classify dedispersed dynamic spectra (which we treat as images) of the FRBs into one of the two sub-classes, i.e., repeater and non-repeater, based on their various temporal and spectral properties and relation between the sub-pulse structures. Additionally, we also used mathematical model representation of the total intensity data to interpret the deep learning model. Upon fine-tuning the pretrained ConvNext on the FRB spectrograms, we were able to achieve high classification metrics while substantially reducing training time and computing power as compared to training a deep learning model from scratch with random weights and biases without any feature extraction ability. Importantly, our results suggest that the morphological differences between CHIME repeating and non-repeating events persist in Catalog 2 and the deep learning model leveraged these differences for classification. The fine-tuned deep learning model can be used for inference, which enables us to predict whether an FRB’s morphology resembles that of repeaters or non-repeaters. Such inferences may become increasingly significant when trained on larger data sets that will exist in the near future.

💡 Research Summary

This paper presents a deep‑learning framework that classifies fast radio bursts (FRBs) into repeating (FRB‑R) and apparently non‑repeating (FRB‑NR) categories using only the morphological information contained in their dynamic spectra. Leveraging the extensive CHIME/FRB Catalog 2 (4 545 events, of which 981 are repeats from 83 sources), the authors extract dedispersed dynamic spectra with the fitburst pipeline, treat each spectrum as a grayscale image, down‑sample the frequency axis by a factor of 64, and resize to 224 × 224 pixels. The grayscale images are duplicated across the three RGB channels and normalized to the ImageNet statistics (mean =