Abstract:Abstract Detecting and mitigating radio frequency interference (RFI) is critical for enabling and maximising the scientific output of radio telescopes. The emergence of machine learning (ML) methods capable of handling large datasets has led to their application in radio astronomy, particularly in RFI detection. Spiking neural networks (SNNs), inspired by biological systems, are well suited for processing spatio-temporal data. This study introduces the first exploratory application of SNNs to an astronomical data processing task, specifically RFI detection. We adapt the nearest latent neighbours (NLNs) algorithm and auto-encoder architecture proposed by previous authors to SNN execution by direct ANN2SNN conversion, enabling simplified downstream RFI detection by sampling the naturally varying latent space from the internal spiking neurons. Our subsequent evaluation aims to determine whether SNNs are viable for future RFI detection schemes. We evaluate detection performance with the simulated HERA telescope and hand-labelled LOFAR observation dataset the original authors provided. We additionally evaluate detection performance with a new MeerKAT-inspired simulation dataset that provides a technical challenge for machine-learnt RFI detection methods. This dataset focuses on satellite-based RFI, an increasingly important class of RFI and is an additional contribution. Our SNN approach remains competitive with the original NLN algorithm and AOFlagger in AUROC, AUPRC, and F1-scores for the HERA dataset but exhibits difficulty in the LOFAR and Tabascal datasets. However, our method maintains this accuracy while completely removing the compute and memory-intense latent sampling step found in NLN. This work demonstrates the viability of SNNs as a promising avenue for ML-based RFI detection in radio telescopes by establishing a minimal performance baseline on traditional and nascent satellite-based RFI sources and is the first work to our knowledge to apply SNNs in astronomy.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the impact of radio - frequency interference (RFI) on radio astronomical observations. As the number, types of RFI sources and the overall interference to observations keep increasing, detecting and mitigating RFI is crucial for modern radio observatories. This paper focuses particularly on the problem of using Spiking Neural Networks (SNNs) to detect RFI. ### Background and Objectives of the Paper 1. **Importance of RFI Detection**: - RFI poses a significant challenge to current and future radio telescopes. - Detecting and mitigating RFI is essential for maximizing the scientific output of radio telescopes. 2. **Limitations of Existing Methods**: - Existing RFI detection algorithms can be divided into analytical methods and machine - learning (ML) - based methods. - Although ML methods are superior to traditional methods in some aspects, they usually require a large amount of labeled training data, which is both expensive and time - consuming. - The high computational cost of ML methods limits their practical applications. 3. **Research Objectives**: - Introduce SNNs into the field of astronomy, especially for RFI detection. - Propose a new RFI detection algorithm, namely the Spiking Nearest Latent Neighbors (SNLNs) method. - Evaluate the performance of SNNs in RFI detection and compare it with existing ANN methods. ### Main Contributions 1. **Introducing SNNs into Astronomy**: - For the first time, SNNs are applied to the field of astronomy, opening up a new research direction. 2. **Proposing the SNLNs Algorithm**: - Based on the unsupervised NLN method, by converting ANN to SNN, the time - varying characteristics of SNN are used to simplify the downstream RFI detection task. 3. **New RFI Detection Dataset**: - Create a new dataset based on the MeerKAT telescope, focusing on simulating RFI caused by satellites, which is an increasingly important category of RFI. 4. **Performance Evaluation**: - Use the HERA, LOFAR and newly created Tabascal datasets to evaluate the performance of the SNLNs method. - Compare with the traditional NLN method and AOFlagger, demonstrating the competitiveness of SNNs in RFI detection. ### Method Overview 1. **Network Architecture**: - Use a two - layer stride auto - encoder, including optional batch normalization and dropout layers for regularization. 2. **Training Strategy**: - Conduct a hyperparameter search for each dataset and optimize the hyperparameters to obtain the highest F1 score. - Use Optuna for hyperparameter optimization, specifically using the default tree - structured Parzen estimator algorithm. 3. **SNN Conversion**: - Use the SpikingJelly library to convert the trained ANN to SNN. - Replace the neuron layer with spiking neurons and adjust the weights to adapt to the LiF neuron implementation. 4. **SNLN Algorithm**: - Input each test patch into the SNN auto - encoder and run for multiple time steps. - Decode the internal latent representation from the last k frames output by the auto - encoder. - Average these output frames and subtract the result from the original test patch to obtain the noise mask. ### Data Selection 1. **HERA Dataset**: - Use Hydrogen Epoch of Reionisation Array simulation data. 2. **LOFAR Dataset**: - Use manually labeled LOFAR observation data. 3. **Tabascal Dataset**: - A newly created dataset, simulating the MeerKAT telescope, focusing on including RFI caused by moving low - orbit satellites. ### Results and Evaluation 1. **Performance Metrics**: - Use AUROC, AUPRC and F1 score to evaluate algorithm performance. - The SNLNs method performs comparably to the original NLN method on the HERA dataset and shows better performance on the LOFAR and Tabascal datasets.

RFI detection with spiking neural networks

RFI Detection with Spiking Neural Networks

Spiking Neural Networks for Radio Frequency Interference Detection in Radio Astronomy

Supervised Radio Frequency Interference Detection with SNNs

A Self-learning Neural Network Approach for Radio Frequency Interference Detection and Removal in Radio Astronomy

Learning to detect RFI in radio astronomy without seeing it

A robust RFI identification for radio interferometry based on a convolutional neural network

A CNN and LSTM-Based Approach to Classifying Transient Radio Frequency Interference

Radio frequency interference detection using efficient multiscale convolutional attention UNet

Radio Frequency Interference Detection Using Efficient Multi-Scale Convolutional Attention UNet

Spiking Neural Networks for Detecting Satellite-Based Internet-of-Things Signals

A comparison framework for deep learning RFI detection algorithms

Deep residual detection of radio frequency interference for FAST

DSC based Dual-Resunet for radio frequency interference identification

Semantic Segmentation of Solar Radio Spikes at Low Frequencies

Spike-based Encoding and Learning of Spectrum Features for Robust Sound Recognition.

Anomaly Detection and RFI Classification with Unsupervised Learning in Narrowband Radio Technosignature Searches

Lightweight deep neural network for radio frequency interference detection and segmentation in synthetic aperture radar

Radio Frequency Interference Detection and Mitigation Using Compressive Statistical Sensing

COSMIC-2 RFI Prediction Model Based on CNN-BiLSTM-Attention for Interference Detection and Location

Remove First Detect Later: a counter-intuitive approach for detecting radio frequency interference in radio sky imagery